Option 1: Transaction Processing Systems 363

In this chapter you will learn to:

• recognise and describe a transaction • identify situations where data warehousing and data
mining would be an advantage
• identify, describe and use a batch transaction
processing system • assess the impact on participants involved in transaction
processing
• distinguish between the storage of collected data
and the storage of processed data in a batch system • identify jobs that have changed and/or jobs that have
been created as a result of transaction processing, and
• identify, describe and use a real time transaction
report on the implications of these changes for
processing system
participants in the system
• compare and contrast batch and real time
• discuss alternatives for when the transaction processing
transaction processing
system is not available and explain why they need to be
• analyse an existing transaction processing system periodically tested
to determine its strengths and weaknesses
• identify security, bias and accuracy problems that could
• design and implement procedures for validating arise from the actions of participants
entered data
• recognise the significance of data quality
• assess the work routine of a clerk in a manual
transaction system to determine its suitability for
automation Which will make you more able to:
• identify participants, data/information and • apply and explain an understanding of the nature and
information technology for the given types of function of information technologies to a specific
transaction processing systems practical situation
• describe the relationships between participants, • explain and justify the way in which information
data/information and information technology for systems relate to information processes in a specific
the given types of transaction processing systems context
• for a scenario diagrammatically represent • analyse and describe a system in terms of the
transaction processing using data flow diagrams information processes involved
• distinguish between the different types of TPS • develop solutions for an identified need which address
• store digital data in databases and other files in all of the information processes
such a way that it can be retrieved, modified and • evaluate and discuss the effect of information systems
further processed on the individual, society and the environment
• implement systems to store paper transactions • demonstrate and explain ethical practice in the use of
• select and apply backup and recovery procedures information systems, technologies and processes
to protect data • propose and justify ways in which information systems
• document, including diagrammatical will meet emerging needs
representations, the steps in batch processing • justify the selection and use of appropriate resources
• document, including diagrammatical and tools to effectively develop and manage projects
representations, steps in real time transaction • assess the ethical implications of selecting and using
processing specific resources and tools, recommends and justifies
• identify systems for which batch is appropriate and the choices
is not appropriate • analyse situations, identify needs, propose and then
• distinguish between on-line real time and batch develop solutions
systems • select, justify and apply methodical approaches to
• create and use a transaction processing system planning, designing or implementing solutions

• describe the operation of relevant hardware and • implement effective management techniques
how each is used to collect data for transaction • use methods to thoroughly document the development
processing of individual or team projects.
• design and justify paper forms to collect data for
batch processing
• design user friendly screens for on-line data
collection
• identify existing procedures that may provide data
for transaction processing
• create user interfaces for on-line real time and
batch updating, and distinguish between them

Information Processes and Technology – The HSC Course

 364 Chapter 4

In this chapter you will learn about:

Characteristics of transaction processing systems • updating in on-line real time systems:
– relevance and impact
• a transaction - a series of events important to an – technology required
organisation that involve a request, an – hardware requirements - large secondary storage
acknowledgement, an action and an outcome – software requirements – on-line database and user
• the components of a transaction processing system, friendly interface
including: – steps in on-line real time processing
– purpose – suitable applications
– data
– information technology Other information processes in transaction processing
– processes systems
– participants • collecting in transaction processing:
• batch transaction processing - the collection and – hardware including
storage of data for processing at a scheduled time or - Automatic Teller Machines (ATM)
when there is sufficient data - barcode readers
- Radio Frequency Identification (RFID) Tags
• real time transaction processing - the immediate – collection from forms
processing of data – screen design for on-line data collection
• the significance of data validation in transaction – web forms for transaction processing (real time and
processing batch)
• the historical significance of transaction processing • analysing data, in which output from transaction
as the first type of information systems processing is input to different types of information
systems, such as:
Types of transaction processing systems
– decision support
• web-based – management information systems
• non web-based – data warehousing systems (for data mining)
• on-line real time – enterprise systems
• batch Issues related to transaction processing systems
• systems that appear real time, responding as the • changing nature of work and the effect on participants,
transactions occur, but where the actual updating is including:
batch processed, such as credit card transactions – the automation of jobs once performed by clerks
– shifting of workload from clerks to members of the
Storing and retrieving in transaction processing public
systems
• the need for alternate procedures to deal with
• storage of digital data in databases and files transactions when the TPS is not available
• retrieval of stored data to conduct further transaction • bias in data collection:
processing such as printing invoices – when establishing the system and deciding what
• systems to store paper records of transactions data to collect
• data backup and recovery, including: – when collecting data
– grandfather, father, son • the importance of data in transaction processing,
– off-site storage including:
– secure on-site storage – data security
– full and partial backups – data integrity
– recovery testing – data quality
– suitable media • control in transaction processing and the implications it
– specialised backup software has for participants in the system
– transaction logs
– documenting backup and recovery procedures • current and emerging trends in transaction processing
– mirroring – data warehousing and data mining
– rollback – Online Analytical Processing (OLAP) and Online
Transaction Processing (OLTP)
• updating in batch systems:
– historical significance
– limitations of batch processing
– technology required
– steps in a batch update
– suitable applications

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 365

4
OPTION 1
TRANSACTION PROCESSING SYSTEMS
Transaction processing systems are crucial to the operation of most finance, banking
and electronic commerce organisations. Transaction processing is primarily concerned
with maintaining data integrity. Such systems can operate at the single database level,
but they also operate at higher levels where data in many databases and even many
different systems is involved. For example transferring funds from one financial
institution to another.
So what is a transaction? A transaction is a Transaction
series of events that when performed A unit of work composed of
together complete some unit of work that multiple events that must all
is important to an organisation. Each succeed or must all fail. Events
transaction has two possible outcomes, perform actions that create
either it is a complete success or it is a and/or modify data.
complete failure.
If a transaction is successful then all the events contained within the transaction must
have performed their actions successfully. However, if one or more events are unable
to complete their actions then the whole transaction must fail, which requires the data
to be left in the same state it was in prior to the transaction commencing. This means
any events that could successfully perform their actions must be stopped. For example
when transferring funds between accounts two events must occur; one account is
debited and another credited. If the debit event fails then the credit event must be
stopped, similarly if the credit event fails then the debit event must be stopped.
Managing the success or failure of transactions is an essential process performed
during transaction processing. Transaction processing systems include mechanisms
for ensuring events can be completed successfully, but not yet permanently.
Essentially the transaction processing system requests that each event occur and
receives a response indicating that the actions performed are guaranteed to succeed or
have failed. If a successful response is received for all events then the transaction as a
whole can be “committed”, meaning each event is requested to store its data changes
permanently within the appropriate databases or systems. If one or more events have
failed then the transaction is “rolled back”, meaning each event is requested to abort
all actions. In response each event sends an acknowledgement to confirm they have
performed the request.
A transaction can include events that perform actions on a single database, many
databases or on a variety of different information systems. These databases and
systems can be widely distributed and in some instances they are operated by different
organisations. The detail of how such transactions are processed will become clearer
throughout the chapter.

GROUP TASK Discussion

Brainstorm a list of typical transactions and their component events and
actions. Discuss problems that may cause these transactions to fail.

Information Processes and Technology – The HSC Course

366 Chapter 4

In this Option we commence by examining characteristics of transaction processing

systems. This includes a brief examination of the history of transaction processing,
how transaction processing automates manual tasks, the components of transaction
processing systems and how such systems maximise the accuracy of data. We then
examine real time and batch transaction processing systems. In real time systems each
transaction is immediately processed online, whilst batch processing collects input
data over time and then at some later time batches of many similar transactions are
processed. Backup and recovery strategies and technologies are examined. We then
examine collection hardware and forms used for collection. The data within
transaction processing systems is used as input to other systems; we briefly consider
examples of such systems. Finally we discuss issues related to transaction processing
systems.
CHARACTERISTICS OF TRANSACTION PROCESSING SYSTEMS
Transaction processing is one of the earliest commercial uses of computer systems. In
this section we examine some early examples of transaction processing to illustrate
how such systems automate and improve upon manual transaction processing. We
then examine features of manual transaction processing systems that make them well
suited to automation. Finally we examine modern transaction systems – their
components and how they maintain the accuracy of data.
HISTORICAL SIGNIFICANCE OF TRANSACTION PROCESSING
The operations performed by transaction processing systems were up until the 1950s
performed solely by clerks using manual processes. Early computers were originally
developed to solve scientific and mathematical problems for government and military.
It was during the 1950s that the application of computers to business and financial
records emerged.
Prior to the 1980s it was common for complete transaction processing applications to
be developed (often in Cobol) for each individual organisation. During the 1980s
database management systems emerged to manage and control access to databases.
Today most transaction processing systems are based on one or more relational
database management systems (RDBMS) with client applications being written to
meet an organisation’s specific needs.
Some of the significant developments
that have led to today’s transaction
processing systems are outlined
below.
• UNIVAC I (Universal Automatic
Computer), released in 1951, was
the first commercially produced
computer to gain wide acceptance
by the public. The UNIVAC I was
based on vacuum tubes and was
the first computer to be routinely
used for batch processing of
Fig 4.1
business transactions. UNIVAC I UNIVAC I the first commercially available
was designed and built by John computer used for transaction processing.
Presper Eckert and John William
Mauchly. Their company Eckert-Mauchly Computer Corporation was bought by
Remington Rand – both Eckert and Mauchly continued to work for Remington
Rand after selling the company. The UNIVAC II and UNIVAC III were
subsequently released.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 367

• The programming language Cobol (Common Business Oriented Language) was

developed in 1959. At the time computers were largely used for scientific and
mathematical calculations. Cobol, as its name suggests, was targeted directly at
business applications and is still widely used on large mini and mainframe
computers. Cobol was the first language for large scale transaction processing.
• In 1964 IBM released its highly successful System/360 range of computers and
peripherals (see Fig 4.2). These general-purpose systems supported approximately
40 different peripherals and included the ability to include redundant components
to improve fault tolerance. Information systems based on System/360 supported
real time input and processing from hundreds of attached terminals.
• In 1969 IBM released the first version of CICS
(Customer Information Control System). In
terms of transaction processing, CICS is a
transaction processing monitor or TPM – it
manages the processing of transactions from
multiple clients to multiple servers. This
software product has been continuously
upgraded and is widely used today.
• SQL was first developed in the early 1970s by
IBM under the name “System R”. The design
ideas for System R were a direct result of Ted
Codd’s work – Codd is considered the founder
of relational database theory. At the time
System R was viewed as a product to allow
users to directly interrogate databases. The
original designers never intended it to become a Fig 4.2
IBM System/360 Model 65 operator
language that would be used from within consol attached to the CPU.
applications.
• In the early 1980s commercial general-purpose relational database management
systems (RDBMS) emerged. Oracle just beating IBM’s release of DB2. These
systems used SQL both to create relational databases using DDL (Data Definition
Language) statements and to view and update relational data using DML (Data
Manipulation Language) statements.
• SQL first became an ANSI (American National Standards Institute) standard in
1986, however most current database systems, although compliant with most of
the standard SQL syntax also include their own non-standard extensions.
• Microsoft entered the RDMS market in the 1990s with its SQL Server product.
Microsoft SQL Server evolved from the Sybase DBMS; Microsoft dissolved their
partnership with Sybase and renamed their product SQL Server.
• Today Oracle, IBM’s DB2 and Microsoft’s SQL Server dominate the market,
however some open source products such as MySQL have significant market
share within small to medium sized organisations.
• Today large enterprises such as banks, large corporations and government
departments use transaction processing monitor (TPM) software to manage
transactions across a variety of databases and applications of different types,
operated by different organisations and in different locations. Common TPMs in
use include IBM’s CICS and Encina products, BEA System’s Tuxedo software
and more recently Microsoft Transaction Server (MTS). TPMs are an example of
enterprise systems as they manage critical data and processes across an entire
organisation.
Information Processes and Technology – The HSC Course
368 Chapter 4

GROUP TASK Discussion

Initially transaction processing software was written for each specific
application. Today it is common to use standard DBMS and TPM
software and only the client applications are custom solutions. Propose
likely reasons why this has occurred.

GROUP TASK Research

Research and identify examples of enterprise systems. Determine whether
these systems process transactions.

AUTOMATION OF MANUAL TRANSACTION PROCESSING

Processing of manual transactions almost always follows a strict sequence of events.
Each event must be acknowledged as complete before the next commences and if any
event fails then the entire transaction is aborted. In manual systems, events are
performed by clerks and other personnel according to strict predefined rules. Indeed in
large organisation’s it is common for each clerk to repetitively perform just one of the
events within each transaction. The transaction is handed onto the next clerk
responsible for the next event in the sequence. The strict sequence and rules of such
transactions make them particularly well suited to automation using computers.
As a simple example let us consider a manual system used within a small store and
then assess the benefits of automating this system. The store is operated by a husband
and wife team who have time during the day to perform all sales, purchasing,
stocktaking and other transactions manually. The store uses a simple cash register,
which is essentially a calculator with an attached cash drawer. The cash register does
keep a total of all sales processed during the day. The store has an EFTPOS terminal,
which operates as a separate system.
Sales – this transaction occurs to process each customer’s purchases.
1. Locate price on item and enter into cash register.
2. Repeat 1 for each product.
3. Calculate total.
4. Receive payment from customer.
5. If EFTPOS payment then wait for approval and hand EFTPOS receipt to customer.
6. Enter payment amount into cash register.
7. If cash payment then calculate and hand change to customer.
8. Hand register receipt to customer.
Stocktake – this transaction collects data to enable the storeowner to calculate the
quantity of each product to purchase.
1. Make a photocopy of stocktake sheets. These sheets specify the required number of
each product when fully stocked, the product’s supplier and also columns for
recording current stock in the store.
2. Count and record number of each product on shelves.
3. Count and record number of each product in storeroom.
Purchasing – this transaction produces purchase orders for each supplier.
1. Complete Stocktake.
2. Calculate number of each item to purchase and record on stocktake sheets.
3. Create purchase order for supplier.
4. Work through stocktake sheets recording each product from current supplier.
5. Calculate order total.
6. Repeat steps 3, 4 and 5 for each supplier.
7. Submit all purchase orders to suppliers via fax.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 369

GROUP TASK Discussion

Three transactions are described above, however there are other
transactions that need to occur. Propose other likely transactions and
outline a set of possible events occurring to perform these transactions.

Notice that much of the data used by all three of the above transactions is the same. It
is the information generated by the transaction that is different. Furthermore the
output from one transaction is used as data for another transaction. For example each
sales transaction reduces the amount of stock, and each stocktake transaction produces
the data required for purchasing. Such observations make this system well suited to
automation. The flow of data and information entering and leaving each of these
transactions is modelled on the data flow diagram in Fig 4.3. Note that each of the
transactions is represented as a process as they are composed of events that process
data in some way. Each of these transactions could be expanded into a lower level
DFD or a step-by-step description that details their component events.
Number of product on
Product shelves and in storeroom Stocktake
Products
Type Stocktake Sheets
in store

Product Type, Supplier,

Product Fully stocked number,
Customers Price Number on shelves and
in storeroom.
Purchasing
Payment Details
Sales
Register Receipt Details Purchase order details Suppliers

Fig 4.3
Data flow diagram modeling the flow of data between a store’s manual transactions.
The stocktake sheets perform many of the tasks performed by a database, hence on the
DFD a data store is used. They store all the data required by the purchasing
transaction process. In addition the stock take sheets allow processing to halt between
the stocktake process and the purchasing process. The Products in Store entity could
also have been represented as a data store as each product stores its price in the form
of a price tag and its product type. In reality these are the actual products, hence
representing them as an external entity makes more sense.
No doubt it is clear that this system could be automated using a relational database to
integrate sales, product, supplier, orders and stocktake data. Later in this chapter we
shall examine a point of sale (POS) system, which is essentially an automated version
of the above manual system. At this stage we are interested in the strengths and
weaknesses of manual systems and of automation. Let us consider some general areas
relevant to most manual systems together with common strengths and weaknesses of
automation. We shall then discuss our local store example in an attempt to assess its
suitability for automation.
Manual system strengths:
• Minimal start-up costs – little or no initial capital expenditure.
• Minimal training time and costs.
• Quick response to changing requirements.
• Well suited to small organisations where participants have time and fulfil multiple
roles.
• Responds well to human insight and intuition.

Information Processes and Technology – The HSC Course

370 Chapter 4

Manual system weaknesses:

• Analysis of historical data is difficult and time consuming.
• Transactions take considerably longer to process.
• Difficult to rigidly enforce transaction rules and sequences.
• Redundant or duplicate data is a feature of most manual systems.
• Some human errors are to be expected.
• System becomes more and more difficult to manage as it grows.
• Making backups of data is difficult and is rarely, if ever, performed for all data.
Automated transaction processing strengths:
• Much faster transaction processing.
• Less repetitive work for participants.
• Enforces the sequence and rules for each transaction.
• Calculation errors are virtually eliminated.
• Ability to integrate transaction processing with outside organisations.
• Historical data available for statistical and financial analysis.
• Backups easily made and restored if system fails.
• System easily grows as transaction processing needs grow.
Weaknesses of automated transaction processing:
• Significant start-up costs to purchase information technology.
• Extensive training required to operate the system.
• Changes to requirements often require specialised expertise to implement.
• Rigidly enforces existing transaction rules and sequences for all data.
• Less total work for humans resulting in lower employment.
• Reliance on information technology – failure of one or components can cripple the
entire system.
In our local store example the storeowners are a husband and wife team who are
currently able to complete the manual transactions. In this case the time saved through
automation is unlikely to result in increased profits. Furthermore the cost required to
set-up and learn to operate a new automated system is unlikely to be justified. If the
storeowners had sufficient expertise to design and develop their own automated
system then this would be worthwhile. Without an automated system it is difficult for
the owners to accurately monitor sales trends over time. If they were able to perform
such historical analysis then perhaps significant savings could be made by
maintaining more efficient stock levels of products according to predicted demand at
different times of the year. It is likely that this is currently occurring in a somewhat
ad-hoc manner for obviously seasonal items – such as ice creams, Christmas
decorations and gloves.

Consider the following:

Each of the following businesses currently use a manual system for recording their
various transactions.
• A hardware store that stocks thousands of different items and has a staff of 8
employees working at all times.
• A small bookstore that is able to supply any title but maintains minimal stock. The
store purchases titles as they are ordered by customers.
• A carpenter who substantially does subcontract work for 3 builders but does do
some small jobs for residential customers.
• An eBay store that started out selling approximately 5 items per day, but is now
selling 50 items per day.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 371

GROUP TASK Discussion

Assess the suitability of an automated transaction processing system for
each of the above businesses. Discuss likely advantage and disadvantages
of automation compared to retaining their existing manual systems.

COMPONENTS OF TRANSACTION PROCESSING SYSTEMS

In this section we examine the various components of transaction processing systems.
Like all information systems (see Fig 4.4), transaction processing systems operate
within an environment that includes information processes, participants,
data/information and information
technology. All these components work Environment Users
together to meet the system’s purpose.
Each transaction is an information Information System Purpose
process and is therefore composed of
events that are also information
processes. For example adding a new Information Processes
customer to a database involves
collecting and storing information Resources
processes. All information processes are
performed using the resources within the Data/ Information
Participants
Information Technology
system. The system’s resources include
participants, data and information, and
information technology. Boundary
We have already introduced the general
Fig 4.4
nature of transactions and much of the Components of an information system.
remainder of this option continues to
examine different types of transactions in more detail. Therefore in this section we
shall concentrate on participants, data/information and information technology within
transaction processing systems. Recall that participants are people who carry out or
initiate information processes. Information technology includes the hardware and
software that carries out information processes.
Participants
Anybody who interacts directly with a transaction processing system becomes a
participant in that system – they are integral to the system’s operation. Therefore
participants include people who work for the organisation that operates the transaction
processing system and also people (often customers of the organisation) who enter
data that initiates transactions. For example a bank employee is a direct user (and
participant) as they initiate the printing of monthly bank statements. Customers
become direct users (and participants) when they use Internet banking to initiate say
the transfer of money between accounts. On the other hand indirect users are not
participants, they send and/or receive data from the system but do not directly cause
its entry or display. For example when a monthly or quarterly bank statement is
received by a customer in the mail the customer is not a participant, rather they are an
indirect user as they did not initiate the generation of the statement directly.
People in the environment only become participants in online real time transactions.
Real time transactions are performed immediately in response to user or participant
input. On the other hand for batch processes people from the environment may well
provide data, but no transaction processing occurs or is initiated by that person. Rather
the transaction is performed along with other similar transactions at some later time or
when a sufficient quantity of data is present. Consider the difference between a
Information Processes and Technology – The HSC Course
372 Chapter 4

transaction performed using an ATM compared with writing a cheque. When using an
ATM the user initiates and therefore causes the transaction to be performed
immediately – essentially they are performing duties similar to a bank teller. However
when writing a cheque the customer has little control over when the transaction is
actually processed and furthermore they are not interacting directly with the bank’s
transaction processing system.
Data/Information
In the majority of transaction processing systems data is stored in databases – usually
relational databases. This data is transformed into information by the system’s
information processes. We studied the organisation and design of relational databases
in depth in chapter 2. All the information in regard to tables, records, relationships,
referential integrity, data validation, data integrity and data verification applies to
transaction processing systems. However in transaction processing systems a further
issue exists – how to ensure the integrity (correctness and accuracy) of data during
transactions. What if another user or process views or alters data during a transaction?
What if the data received from another system has problems? What if the system fails
in some way during a transaction? In regard to data and information such issues are
resolved by recording the detail of all transactions in a transaction file or log. How
these transaction records help to resolve these issues will become clearer in the next
section on data validation and data integrity.
Within transaction processing systems additional data is always created to record
details of each transaction that occurs. In older systems the actual live data was
commonly known as the master file and the details of each transaction was recorded
in a transaction file. The application controlled and managed both the transaction file
and the master file. All changes being recorded in the transaction file during
transaction processing, with changes to the master file only being made when
transactions are finally committed. Newer systems still create such transaction data
(often called a transaction log), however management of this transaction data is left up
to the DBMS and, if used, the transaction processing monitor rather than the
application software. Most commercial operating systems also provide transaction
capabilities as part of the file system.
Such operating systems create transaction Read Record
records that allow actions on complete
files to form part of transactions. These Modify Record
operating system capabilities are also
available to other applications, including Store Modified
transaction processing monitors. Record
To simplify our discussion let us refer to Master file
(or Database)
the transaction data or transaction file as a
transaction log and the actual data as the Transaction Commit?
log
master file. Recall that transactions can be
committed or rolled back. The transaction Yes
log contains the essential data that Read Record
facilitates this ability. When an event
occurs as part of a transaction two Overwrite
possibilities arise: Record
1. Fig 4.5 describes the first possibility
for an event that modifies a single Fig 4.5
record. The event occurs however the Flowchart describing modifying a record as part
of a transaction where the master is not altered
changed or added records are recorded
until the transaction commits.
in the transaction log and no change is
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 373

yet made to the actual data in the master file. If the transaction is committed then
the records in the transaction log replace or are added to the master file. If the
transaction is rolled back then the records in the transaction log are not written to
the master file.
2. Copies of the original unchanged data are recorded in the transaction log and then
the changes are made immediately to the actual data within the master file. If the
transaction is committed then nothing more needs to occur. If the transaction is
rolled back then the record in the transaction file is copied back over the actual
data in the master file. When new records are created as part of a transaction the
transaction log must contain an entry specifying the record to delete should the
transaction be rolled back.
GROUP TASK Activity
Create a flowchart to model the processes occurring to modify an existing
record in the master file using the second strategy described above.

In either case the transaction log is used to enable the committing or rolling back (or
even rolling forward) of events within transactions. Most current DBMSs actually
record both before and after versions of the data within their transaction logs – in
essence they allow implementation of both the above possibilities. This means the
transaction log is really a log of all the activities performed on the data.
The most compelling reason for maintaining before and after versions of all data
changes is to provide a backup of all recent changes since the last backup. The
database (or master file) can be restored from the most recent backup and then the
transaction file can be used to commit (or roll forward) all transactions performed
since the restored backup was made. If at the time of failure some transactions were
incomplete then those events that formed part of such transactions can be rolled back.
Such restore operations are essentially automated within most modern DBMS and
transaction processing monitor software products.
A complete transaction log is also useful during audits as it shows when, what and
who performed each transaction. Utilities are available for most DBMS products that
allow the transaction log to be analysed in detail. Such utilities also allow transactions
in the log to be rolled back and rolled forward individually.
GROUP TASK Research
Transaction log files continually grow in size – sometimes their size can
exceed the size of the actual database. Research techniques and strategies
for ensuring transaction logs do not grow excessively.

Information Technology
The hardware and software forms the information technology of the system.
Transaction processing systems vary enormously in both size and scope. A small
database may serve just a few local users, however a similarly small database may
serve many more users via the web. Larger critical transaction processing systems
perform thousands or even millions of transactions daily. The hardware and software
requirements vary enormously; hence in this section we shall introduce some general
areas for consideration. Later in this chapter we examine more specific examples
where the detail of the hardware and software can be specified more precisely.
• Hardware
Possible hardware for transaction processing systems includes:
• Server machines that include redundant components to improve fault tolerance.
In medium to large systems multiple servers provide access to the same
Information Processes and Technology – The HSC Course
374 Chapter 4

database. If one server fails or is taken offline for maintenance or upgrading

then the other servers automatically take up the extra load. Larger government
and multinational organisations commonly use mainframe computers that are
able to support thousands of users and access to enormous databases.
• Storage devices with sufficient capacity and data access speeds to support the
size of the database and the number of users. Commonly RAID storage is used.
For transaction logs even on relatively small systems a mirrored RAID solution
is common to ensure that a single failure of one drive does not result in loss of
incomplete and recent transactions.
• Communication devices and transmission media able to support the number of
required users and data access speeds to ensure acceptable response times.
Many servers include multiple NICs to achieve higher data access speeds.
• Backup devices such as tape drives, tape libraries, CD burners and DVD
burners. In some systems complete copies of the data are maintained on
mirrored hard disks located in a different location to the operational data.
• Client workstations for running the client applications and interacting with the
system’s participants and users. These machines may include specialised
collection devices such as barcode scanners, RFID readers, touch screens and
magnetic ink character recognition (MICR) readers. The client machines may be
dedicated devices such as ATMs or EFTPOS terminals or they may be personal
computers connected via the Internet or the organisation’s network.
• Software
Possible software for transaction processing systems includes:
• DBMS software to manage and control the transactions performed on linked
databases. We discussed the operation of DBMSs in chapter 2. The DBMS runs
on one or more servers and provides services to client applications to enable
them to access the databases within the transaction processing system. Each
DBMS includes the ability to manage transactions performed on databases
under its control and includes a transaction log.
• Client applications that are installed on client workstations and provide the
interface for participants to initiate and perform the system’s information
processes. The client applications make requests, often in the form of SQL
statements, to the database servers using a client-server architecture. For larger
systems that perform transactions across many servers or many systems the
client applications send their requests via the transaction processing monitor.
• Proprietary software applications that are designed and developed to meet the
needs of a specific organisation. Such software is common in large transaction
processing systems running on mainframe machines. These software
applications are written from the ground up – including providing, specifying
and controlling access to data directly without the use of a DBMS. The term
proprietary means the software is produced for a specific system or organisation
and is generally owned by that organisation.
• Transaction processing monitors (TPMs) are software applications that
coordinate the transaction processing of large transaction processing systems.
Fig 4.6 describes the general software architecture of transaction processing
systems that include transaction processing monitors. These large systems
commonly include many database servers that may access the same logical
database or may access different databases. In addition transaction processing
monitoring applications can connect to systems operated by other organisations.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 375

Each server or system has its own resource manager (refer Fig 4.6) that makes
available resources to the TPM. A resource manager is essentially a software
product that provides an interface between the resource and the transaction
processing monitor.

Resource
Manager
DBMS
Client
Application
Database

Transaction Processing Monitor

Resource
Manager
Client DBMS
Application

Resource
Manager
Client DBMS Database
Application

Resource
Manager
Client Other
Application System

Fig 4.6
General architecture of a system that includes a transaction processing monitor.

The main task of transaction processing monitors is to ensure the integrity of

transactions that include events which execute on different servers and/or
systems. The TPM controls the commit and rollback of the total transaction in
response to requests sent to and acknowledgements received from resource
managers. Each server or system performs its own lower level transaction and
reports the outcome (success or failure) back to its resource manager, who in
turn communicates with the TPM. In addition TPMs are able to balance the load
of transactions sent to each server. Transaction processing monitors are also
known as transaction managers or transaction processing services. Examples
include IBM’s CICS and Encina products, BEA System’s Tuxedo and
Microsoft Transaction Server.
GROUP TASK Discussion
A user initiates the transfer of funds from a Commonwealth Bank account
to a Westpac bank account. Discuss the role of transaction processing
monitors during this transfer.

DATA INTEGRITY
The integrity of data is critical in all
transaction processing systems. Recall Data Integrity
from our earlier work on database systems A measure of how correct and
(Chapter 2) that data integrity is a measure accurately data reflects its
of how correct and accurate data is source. The quality of the data.
compared to its source. In Chapter 2 we
considered three techniques for improving data integrity, namely data validation, data
verification and also referential integrity. In this section we briefly discuss examples
of each technique within transaction processing systems. We then introduce the ACID
properties of transactions and the type of problems they solve.

Information Processes and Technology – The HSC Course

376 Chapter 4

Data Validation
Data validation checks ensure reasonable data enters the system. In transaction
processing systems data that is incorrect at the time of collection is likely to cause a
variety of different problems when it is later used as part of transactions. There are
two different types of data validation
commonly performed. The first ensures Data Validation
the data entered is of the correct data type A check, at the time of data
and format. This is generally performed collection, to ensure the data is
by the client application. The second is reasonable and meets certain
more difficult as it aims to ensure the data criteria.
entered is correct in terms of the business
rules of the enterprise. That is, it determines if the data is correct in terms of its ability
to be processed. For example when ordering a book the ISBN is often entered as a
unique identifier. Data validation within the client application ensures the correct
number of digits are entered. The book store’s business rules require that the ISBN
must exist within their database. Therefore a query must be executed to validate that
this is indeed true.
A single data entry error that is undetected can affect numerous transactions across
many organisations. For example consider a BPay reference number on a supplier’s
invoice that is being paid by a customer using Internet banking. Let us assume this
invoice must be paid before the goods are shipped. If the BPay reference number is
entered incorrectly by the customer then the total transaction will eventually fail. The
consequences of this simple data entry error is costly for both the customer and also
for the organisations involved in the transaction. The bank must inform the customer
of the problem, however the customer is not aware of any potential problem and
hence they are unlikely to check their bank messages. The supplier does not receive
the funds and therefore will reissue the invoice or simply not supply the goods. The
customer is not happy as they are unaware of the error and hence wonder why their
goods do not arrive. Resolving the problem involves further time and cost for all
parties. These issues could be resolved by validating the BPay reference number prior
to the transaction commencing.
Data Verification
Data verification is used to maintain the integrity of data over time. This is a difficult
task in most information systems and is rarely 100% successful. For example people
and also businesses move location, change their phone numbers, credit card numbers
and even change their names. Ensuring that such changes are reflected in the data is
the aim of data verification processes.
In large government and commercial Data Verification
transaction processing systems data A check to ensure the data
collected and stored matches
verification becomes an enormous
and continues to match the
undertaking. Currently in Australia there is
source of the data.
no single unique identifier that can legally
be used to identify individuals across all
these systems. If such an identifier was available then it would be possible for
individuals and organisations to change their details in one place and have these
changes replicated to other systems. Privacy concerns prevent such practices. For
example in the mid 1980s the federal government attempted to release the “Australia
Card”, which was to contain a unique number for each Australian citizen and resident.
This number was to be used to link records between most government departments
and even between commercial organisations. As a result of public outcry over privacy
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 377

concerns the legislation was never passed. Currently tax file numbers (TFNs) and
Australian Business Numbers (ABNs) are shared between many government agencies
albeit with strict controls in regard to how data can be linked and used. In Australia it
is illegal for private organisations to use TFNs and ABNs to link data from multiple
sources.
GROUP TASK Discussion
Discuss advantages and disadvantages of widespread use of a unique
identifier for each Australian citizen and resident.

Referential Integrity
In relational databases referential integrity ensures all foreign keys in linked tables
match a primary key in the related table. This means a record in the primary table
must exist before records can be added to the table containing the linked data. If
referential integrity is not enforced then orphaned records will exist. In general such
records cause significant problems when queries are executed on the database.
Within a single database referential integrity is enforceable and hence problems
simply cannot occur within the database. When many databases are involved or
identifiers are being entered by users then problems are inevitable. Data validation
and verification issues can affect referential integrity. For instance, entering an
incorrect BPay reference number means that the primary records held in the various
organisations’ databases cannot be linked to the customer’s payment. The Australia
Card aimed to provide a primary key for each Australian that could be used as the
foreign key in many linked databases. Both systems are attempting to use a unique
identifier in an attempt to enforce referential integrity across multiple databases.
GROUP TASK Discussion
Brainstorm real world examples of data validation and data verification
that aim to improve the referential integrity (and therefore the data
integrity) of databases.

ACID Properties
ACID is an acronym for atomicity, consistency, isolation and durability. The aim is to
ensure all transactions comply with these four properties. They ensure that
transactions are never incomplete (atomicity), the data is never inconsistent
(consistency), transactions do not intrude or affect each other (isolation) and that the
results of a completed transaction are permanent (durability). All these properties
combine to ensure the integrity of all data is maintained before, during and after each
transaction.
To illustrate each of the ACID properties let us use an example transaction – making
an airline reservation using a credit card. This transaction includes the following
general sequence of events:
1. Reserve a seat on a specific flight.
2. Process and approve credit card payment.
3. Issue and record ticket details.
• Atomicity
To be atomic all events within a transaction must complete successfully or none at all.
If any single operation fails then the entire transaction is aborted. This involves rolling
back all events completely so that the data is returned to its original state. If all events
are successful then the transaction is committed, which means the data changes are
made permanent or durable.

Information Processes and Technology – The HSC Course

378 Chapter 4

In our airline transaction imagine what would occur if just one operation failed but the
others were committed. If no seat were reserved then the passenger would arrive with
a paid ticket but with no available seat. If the payment is not processed and approved
then the passenger receives a seat and ticket for free – great for the passenger, but not
so good for the airline. If no ticket is issued or recorded then the passenger and airline
have no record of the transaction resulting in the passenger being refused a seat.
• Consistency
The consistency property ensures transactions take data from one consistent state and
then when the transaction completes the data is left in a consistent state. For a single
event on a single database this is enforced using referential integrity and validation
rules. When the transaction includes many events and spans many databases or
systems then consistency must apply across all these databases and systems.
In our airline transaction a business rule is likely to require the total number of
reserved seats to be equal to the number of tickets issued. If a seat is reserved but does
not result in a ticket being issued then the data is inconsistent in regard to this
business rule. Many other rules are also likely, such as, a customer must be assigned
to each reserved seat, all tickets must be paid for and each ticket must be assigned to a
specific flight and passenger.
• Isolation
Transactions must process data without interfering with or being influenced by other
transactions that are currently executing. In effect each transaction logically executes
in isolation to all other transactions. During the processing of a transaction the data is
often placed in an inconsistent state. For example when transferring funds between
accounts, money is debited from one account and then credited to another account.
After the debit but before the credit the data is in an inconsistent state. This
inconsistent state should not be exposed to other transactions. Furthermore the records
involved should not be available for other transactions to change until the transaction
is completed. If the isolation property is not observed then queries will return
inconsistent results and other transactions will process with potentially erroneous data.
In small systems where only one transaction executes at a time the isolation property
is simple to achieve as one transaction completes before the next commences. If many
transactions can execute at the same time then the solution is more involved. However
even the largest transaction processing systems must ensure their method of
implementing the isolation property results in the same effect as executing each
transaction sequentially.
When multiple transactions can execute concurrently all data involved in a transaction
must be locked such that other transaction processes cannot alter it. We discussed
record locking strategies used by DBMSs in chapter 2 – these strategies are also used
within transactions that span multiple databases and systems. Note that locking does
not alter the actual data, rather it prevents other operations from changing the data. As
a transaction is committed the actual data is altered. Significantly other processes are
aware that a record has been locked by another transaction. Therefore other
transactions must wait for the lock to be released before they proceed.
Record locking, transaction logs and the “two-phase commit” nature of transactions
all influence each other and combine to implement the isolation property. The term
“two-phase commit” refers to events being performed temporarily (phase one) during
a transaction and then being committed (phase two) if the transaction completes
successfully. The first phase is recorded in the transaction log and also involves the
record being locked. The second phase alters the actual data permanently and releases
the record lock.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 379

Consider our airline transaction example. Imagine the isolation property is not present
and a single seat remains available on a flight. Many passengers are now able to
simultaneously reserve this single seat successfully, as long as each transaction
commences prior to committing one of the other transactions. Furthermore they will
go on to pay and be issued with a ticket. When passengers board the flight the airline
will discover there are more passengers than available seats.
• Durability
Durability ensures that committed transactions are absolutely permanent.
Theoretically this means that even if the whole world crashes the changes made by the
transaction will be OK. In real systems durability ensures that during a commit the
results are actually written to some physical storage device. Notification of a
successful commit can therefore be reasonably relied upon.
At first it may seem that executing an update query when committing will ensure
durability of the changes, however in many systems data is held in RAM for a period
of time and is only written to secondary storage as required. Such systems improve
performance, however if power is lost then the contents of RAM is permanently lost.
Therefore durability specifically requires all changes to be written to permanent or
secondary storage before the transaction is truly committed.
In our airline example, imagine an example transaction is apparently committed
successfully. Now say the durability property is not present within the “issuing and
recording ticket” event. Suppose the system fails and this operation is not recorded.
When the passenger goes to board the flight their ticket will not exist on the system.
However inconsistencies will be present as a reservation will exist for the passenger
and a record of payment also exists. Resolving this issue will be costly in terms of
time and also in terms of inconvenience for the passenger.

HSC style question:

Define the term transaction and explain how data integrity is maintained during
processing of transactions.
Suggested Solution
A transaction is a unit of work composed of a sequence of events. All actions
performed by all events must succeed for the transaction to be committed
permanently. If any single event within a transaction fails then all events within the
transaction are aborted or rolled back. Commonly each event within a transaction
alters data within a database.
Whenever data is altered the potential exists for inaccuracies to be introduced and the
integrity of data to suffer. Transactions avoid such possibilities through their ACID
properties. Atomicity ensures a transaction succeeds completely or fails completely.
Consistency ensures each transaction takes the data from one consistent or correct
state to another consistent or correct state. This means inaccuracies or data integrity
issues are only possible during processing of a transaction. This possibility is dealt
with by the isolation property. This ensures data changes are not available to other
transactions until they have been committed. The durability property ensures all
changes made by all events occurring within all committed transactions are
permanently written to storage. This increases data integrity as it guarantees the
consistency of the data after each transaction completes is maintained permanently.

Information Processes and Technology – The HSC Course

380 Chapter 4

SET 4A
1. Which of the following best describes a 6. Examples of TPMs include:
transaction? (A) SQL Server, Oracle, DB2
(A) An event that alters or creates a record (B) SQL Server, CICS, MTS
within a database. (C) Oracle, Encina, Tuxedo
(B) Multiple events that must all succeed or (D) CICS, Tuxedo, MTS
all must fail. 7. The data needed during commit and rollback
(C) A system that controls the execution of processes is stored within the:
many transactions across many (A) transaction log ..
databases or systems. (B) master file.
(D) A process that alters data in different (C) operational database.
records, databases or systems. (D) data source.
2. Transaction processing using computers first 8. Which of the following is the most
emerged during the: significant task performed by TPMs?
(A) 1980s (A) Manage access to many remote DBMS
(B) 1970s servers within an enterprise system.
(C) 1960s (B) Provide an interface between client
(D) 1950s applications and resource managers.
3. A transaction log contains: (C) Manage and control transactions whose
(A) details of the data added or updated events span multiple databases and/or
during processing of transactions. systems.
(B) details of the original data prior to it (D) Force all events within a transaction to
being updated by transactions. be permanently committed.
(C) sequential copies of the data within the 9. Over time existing data becomes less and
master file. less accurate. Which of the following is
(D) Answer A and/or B undertaken to improve this situation?
4. Manual transactions performed by clerks are (A) Data verification.
often well suited to automation because they: (B) Data validation.
(A) are boring and repetitious for (C) Referential integrity checks.
participants to perform. (D) Ensure transactions adhere to the ACID
(B) follow a strict predefined sequence of properties.
rules. 10. Transaction A reads data whilst transaction
(C) can be performed as batch processes. B is executing. Transaction B is rolled back,
(D) commonly include just one operation however transaction A commits. It is later
that alters data. determined that transaction A has introduced
5. Bank customers become participants when inconsistencies into the data. Which ACID
they: property is NOT present?
(A) write a cheque. (A) Atomicity
(B) receive a statement in the mail. (B) Consistency
(C) withdraw cash from an ATM. (C) Isolation
(D) All of the above. (D) Durability

11. Define each of the following terms?

(a) Transaction (c) TPM (e) Data validation
(b) Data verification (d) Referential integrity (f) Data integrity
12. Recommend suitable data validation techniques when collecting each of the following?
(a) Exam marks that are out of 100.
(b) A pair of dates, where the first date must be prior to the current data and the second must be at
least 1 week after the first date.
(c) Adding a product and required quantity to a customer’s order.
13. Outline the history of computer based transaction processing systems.
14. Explain how transaction processing systems implement the ability to commit or rollback
transactions.
15. (a) Why are most manual transactions well suited to automation? Discuss.
(b) What data integrity issues are resolved when all ACID properties are enforced? Discuss.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 381

REAL TIME (ON-LINE) TRANSACTION PROCESSING

Real time transaction processing systems complete transactions immediately they
have been initiated. In most examples of real time transactions an online user initiates
the transaction. Online users include employees of the organisation and also
customers entering details via the web or other networks. Note that when a user
directly interacts (participates) with the system they become a participant in the system.
Real time transaction processing in many references is known as online transaction
processing or OLTP. In real time or online systems, as opposed to batch systems, each
transaction must complete within a reasonable amount of time. If an operation takes
longer than a second or two then feedback must be provided to assure the user that
processing is indeed taking place. Wait times of more than a few seconds are likely to
be aborted by users in the belief that an error has occurred – this is particularly true of
users who are not members of the organisation. This presents significant problems
when transactions are initiated over the web. The organisation has no control over the
speed of transmission once packets reach the Internet. However they do have control
over the speed of data access and processing performed by their hardware and
software. In general, real time transaction processing systems require faster direct (or
random) access to secondary storage, faster and more secure communication links and
more processing power than batch processing systems.
We mentioned above the need for fast response times when using online data entry
forms. We shall discuss the design of such forms in some detail later in this chapter.
However it is worth introducing some often used strategies. Commonly online forms
collect data that is then validated or used as criteria to search a remote database. If the
validation involves simple data type or format issues then it can occur within the data
entry form by the client. If validation involves referring back to the remote database
then response times become an issue. For example when applying for car insurance a
form collects details of the make and model as well as personal and payment details.
To maximise response times it is common for data collection to be split into a
sequence of forms. The data collected on each form being validated or used for a
search prior to display of the next form. In our car insurance example a form is used
to collect the make of car, say Holden. The “make” is used as the search criteria so
that the next form sent to the user need only contain different models of Holden –
Commodore, Berlina, and so on. In addition to reducing the amount of data that needs
to be transferred, sequences of forms also mean just a few user inputs need to be
validated at a time. This improves user friendliness as problems are identified soon
after the user made the input.

Consider the following

Consider a single form that collects say twenty data items over the web. Say two
validation problems are found. The user is then presented with these two items along
with some messages outlining the nature of the problem. Firstly, it will take longer for
the result to be returned to the user and furthermore it may have been many minutes
since the user made the problem entries. Consequently the user is forced to readjust
their thinking to make the corrections. If just a few items are input then validation
messages are returned to the user before their thoughts have moved on.
GROUP TASK Discussion
Are the issues discussed above relevant to online data collected over a
LAN from participants who are part of an organisation? Discuss.

Information Processes and Technology – The HSC Course

382 Chapter 4

In this section we examine three examples of real time (online) transaction processing
systems. We examine a reservation system, and two non-web based systems, namely a
point of sale system and a library loans system. In each case we will identify the
participants, data/information and information technology within the system. We also
model some of the information processes performed as part of the system’s
transactions using data flow diagrams and other system models.
RESERVATION SYSTEMS
Reservation systems are used to collect and process bookings for a variety services.
Examples include hotel, motor vehicle rental, airline and concert reservations. These
systems, although different in terms of the detail of how they are implemented,
process similar transactions. In general a transaction that reserves a service is
composed of the following operations (or information processes).
1. Collect and store details of required service.
2. Confirm availability of service and temporarily reserve service.
3. Collect and store customer details.
4. Collect payment details as required.
5. Process and store payment as required.
6. If successful then commit reservation permanently.
7. Create and display confirmation to customer.
Today many systems allow customers to initiate the processing of reservations via the
web. Many of these systems still provide phone information services or operator
assisted telephone services. The essential processes within each transaction remain
similar regardless of the interface used to communicate with the customer. If payment
or a deposit is required at the time of reservation then it is common for a separate
system operated by a financial institution to be used to process and approve payments.
A typical context diagram for a reservation system is reproduced in Fig 4.7. On this
context diagram participants who work for the organisation are not included as an
external entity. This is clearly correct when customers enter data directly via the
Internet and it is also correct when data is entered into the computer system by
employees of the organisation. In all cases the data originates from the customer
whether they enter data directly via the web (as a participant and user) or via an
operator (as an indirect user).
Service Details,
Customer Details,
Customer Payment Details,
Payment Details
Provider Account Details
Reservation Financial
Customer
System Institution

Service Confirmation Payment Approval

Fig 4.7
Context diagram for a typical reservation system.

There are significant advantages of real time processing of reservations. Customers

are provided with a confirmed booking virtually immediately. The service provider is
able to allocate their service on a first in first served basis and furthermore they have
immediate access to the number of allocated seats, vehicles or rooms on any particular
day or for any particular event. This information enables the service provider to tailor
their marketing and prices to attract customers to poorly performing days or events.
Let us consider the seven steps outlined above in terms of our understanding of
transactions. Do all seven steps need to be successful for the transaction to complete
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 383

or be committed? Yes, if any step fails then all processes should be reversed or rolled
back. Is the sequence of steps significant? Yes it makes logical sense to ensure
availability prior to collecting personal details. It also makes sense to collect personal
details prior to processing the customer’s payment. Clearly confirmation should not
be made until all steps complete correctly.

Consider the following web-based hotel reservation system:

The Hytton is a large city Hotel with a total of 500 rooms. There are four room types,
namely, double, queen double, deluxe double and penthouse in ascending size,
features and price order. The Hotel has rooms on 13 of its 15 floors. Some rooms have
harbour views and generally the view is better on higher floors. Currently the Hotel
charges according to room type without regard to view or floor. Rooms with better
views are assigned to repeat customers and also based on customer requests.
There are numerous different transactions that occur before, during and after a typical
stay at the Hytton. Clearly transactions occur when guests check in, check out, order
movies, food and drinks and also as part of routine operations such as cleaning rooms,
ordering supplies, payroll, and so on. For our purpose we will restrict our discussion
to reserving a room and checking into the hotel upon arrival.
Reservation Transaction (phone-based)
The operations performed during a typical phone reservation transaction are modelled
on the systems flowchart in Fig 4.8. Notice that the Hytton does not require payment
or a deposit at the time a phone reservation is made. The Hotel Database is included
twice on the model simply to improve readability.
Operator answers Enter Search for
phone Guest Name Guest

No Repeat Yes
Hotel Guest?
Database

Create New Retrieve

Guest Record Guest Record

Enter Dates, Verify Guest

Number of Guests, Details
Room Type

Hotel
Calculate available Database
rooms

No Room Yes Create Availability Verify reservation

available? Record details with guest

Fig 4.8
The steps performed for a phone reservation transaction at the Hytton Hotel.

Information Processes and Technology – The HSC Course

384 Chapter 4

GROUP TASK Discussion

Work through the above flowchart (Fig 4.8). Is it true that all the processes
described must complete successfully for the total transaction to succeed?
Discuss.

Reservation Transaction (web-based)

Guests are able to reserve and pay for rooms on-line using the sequence of web forms
reproduced in Fig 4.9. The final confirmation form is not reproduced; it displays a
Reservation Number together with all the confirmed reservation details in a format
suitable for printing. A copy of this confirmation form is also emailed to the guest.

Fig 4.9
Web-based forms for Hytton Hotel reservation transaction.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 385

GROUP TASK Discussion

Analyse the web forms in Fig 4.9 in terms of their design, the purpose of
each of the input fields and the order in which the inputs are collected.

GROUP TASK Discussion

Predict likely processes that are being performed each time the user clicks
the continue or submit button on each of the Fig 4.9 forms. Justify your
predictions using evidence on the forms.

GROUP TASK Discussion

Compare and contrast making a phone reservation with making a web
based reservation from the perspective of a guest.

Check-in Transaction
Guests check-in at the front desk upon arrival at the Hytton Hotel. The following
processes or operations are performed by one of the front desk staff during a typical
check-in transaction:
1. Welcome and determine guest’s name.
2. Find reservation records – guest record and associated availability record.
3. Complete personal details of guest record – address, phone, other guest names.
3. Determine any specific guest requests in regard to view or floor.
4. Assign specific room to guest, which is stored in availability record. Repeat
guests are automatically assigned an available room with the best view.
5. Determine payment method – credit card preferred.
6. If cash or EFTPOS and has not paid then collect and process deposit.
7. If credit card and has not paid in advance then reserve funds for the cost of the
room via EFTPOS terminal.
8. Create charge record for deposit, reserved funds or prepayment.
9. Generate electronic swipe card room key.
10. Print check-in details, attach charge receipt and staple to inside of information kit.
11. Hand information kit and swipe card key to guest and verbally verify all details.
12. Arrange porter to deliver luggage to room.
GROUP TASK Discussion
Identify the participants and also the tasks they perform during the phone
reservation, web reservation and check-in transactions.

Data/Information
The data for the Hytton Hotel system is stored within a relational database. This
database includes the tables and relationships described in the schema shown in Fig
4.10. Note there are many more tables that form part of the complete system – only
those tables used during the phone reservation and check-in transactions are shown.

GROUP TASK Activity

Refer back to the details of the phone/web reservation and check-in
transactions. Determine the records that are accessed, changed and/or
created during these transactions.

Information Processes and Technology – The HSC Course

386 Chapter 4

Charges
Guests ChargeID
1 RoomTypes
GuestID m Date/Time 1
GuestID 1 RoomTypeID
FirstName 1
Description Description
LastName
Charge Notes
PhoneNumber
FaxNumber Availability
Email AvailabilityID Rooms
Address m GuestID 1
RoomID
City m m
StayDate RoomTypeID
Postcode RoomTypeID m Floor
Country RoomID ViewRating
Preferences VingCardID

Fig 4.10
Partial schema for the Hytton Hotel database.
Information Technology
The Hytton Hotel’s system uses a client-server architecture with the database stored
on a RAID storage device attached to the database server. Throughout the Hotel there
are a total of 65 workstations with different hardware configurations. Details of the
hardware and software include:
• The web and DBMS server software runs on separate Dell PowerEdge 2950
Servers. The servers includes two Intel Dual-Core processors and 32GB of RAM.
• The database is managed by Microsoft’s SQL Server DBMS software together
with a customised server application.
• Microsoft’s Internet Information Services web server software runs on the web
server. The web server connects to the Internet via a bank of cable modems. The
cable connections also supply pay television and Internet access to guest rooms.
• Although the Hotel’s website uses SSL, payments are not processed in house. All
online credit card payments are directed to the hotel’s bank where they are
approved and the funds are deposited directly into the hotel’s account.
• The RAID device includes 8 hard disks with a total storage capacity of
approximately 5TB. The system uses RAID 5, which includes both striping and
mirroring to improve both data access performance and fault tolerance.
• The client application within the Hotel
runs on each of the 65 workstations
and has been customised to suit the
particular needs of the Hotel. The
client and server applications are
based on a proprietary hospitality
application.
• The Hotel uses the VingCard security
lock system (see Fig 4.11). Each lock
has its own unique ID and includes
flash memory to store the last 600
entry and exit events. Hotel staff
requiring access to rooms are issued Fig 4.11
with swipe cards, however these do Generating a guest swipe room card using
not operate locks on occupied rooms. the VingCard 2800 terminal.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 387

Swipe cards are coded to operate locks in elevators, hotel entrance doors, other
hotel facilities such as conference rooms and pools, and of course guest rooms.
• 20 laser printers and 7 small receipt printers are installed throughout the Hotel.
• Each workstation runs Microsoft Windows Vista and includes a 100Mbps
Ethernet connection back to the central rack of switches.
• The server connects to a rack of patch panels and Ethernet switches via an Optical
Gigabit interface and cable. Connections to all workstations are cabled using Cat
5e UTP.
• Partial backups are performed each night and full backups each week. Backups are
written to a small attached tape library capable of auto loading 8 tape cartridges
from its built in magazine. Each tape stores 400GB, so total capacity without
manual intervention is approximately 3.2TB.
GROUP TASK Discuss
Briefly explain the purpose of each of the hardware and software items
listed above in terms of performing reservation and check-in transactions.

POINT OF SALE (POS) SYSTEMS

Point of sale systems process transactions within retail outlets. Retail outlets include
small local stores, chains of stores, hotels and clubs, and also large department stores
and supermarkets. Although the amount of data increases significantly for larger
retailers the general nature of the essential transactions remains similar. Retail stores
sell directly to customers and they purchase inventory or stock from suppliers as
described on the Fig 4.12 context diagram. These two processes, purchasing and
selling, form the basic transactions performed by all retailers and hence must be
present within all POS systems. The retailer sends suppliers a purchase order, the
supplier sends the products together with a delivery docket followed shortly by an
invoice. Finally the retailer pays the invoice and sends the payment details in the form
of a remittance advice. When selling retailers accept payment from customers and
provide the customer with a receipt – often in the form of a tax invoice. In addition to
these basic transactions most POS systems provide a variety of additional features to
enhance security, monitor business performance and assist with marketing.
Delivery Docket,
Invoice UPC
Suppliers POS Receipt Details Products
Purchase Order, System in store
Remittance Advice Payment Details Customers

Fig 4.12
Context diagram for a typical POS system.

Particular companies produce and market proprietary POS systems for specific
industries. Some companies produce and market complete POS systems for jewellery
stores, others specialise in hardware stores, whilst others specialise in fruit and
vegetable stores. Commonly these systems include all hardware and software,
together with the training required to operate the system.

GROUP TASK Research

Research examples of proprietary POS systems used by various industries.
Identify features of the POS systems specific to the industry they serve.

Information Processes and Technology – The HSC Course

388 Chapter 4

Most packaged products include a printed Universal Product Code (UPC). Each UPC
is a 12-digit number usually printed with an equivalent barcode on the product’s
packaging. The first 6 digits uniquely identify the manufacturer, the next 5 digits
uniquely identify each of the manufacturer’s products and the final digit is a check
digit. For high value items, such as jewellery, a unique identifier and associated
barcode is commonly created for individual items by the POS software. For products
sold by weight, such as fruit and vegetables, product codes, if used, are added in store
once the product has been weighed and packaged.
GROUP TASK Activity
UPC check digits are calculated by summing the 6 digits in odd positions
and multiplying by three. This result is added to the sum of the 5 digits in
even positions. The check (twelfth) digit is the difference between this
total and the next multiple of ten. Examine UPCs on a number of
products and confirm the check digit is correct.

Let us expand on our initial POS system context diagram in Fig 4.12 by considering
some typical transactions performed by most POS systems.
Sales –this transaction processes each customer’s purchases using the POS terminal.
1. Scan UPC on product packaging.
2. System retrieves product description, price and stock level from database.
3. Stock level reduced by one and stored in database.
4. Repeat 1 to 3 for each product.
5. System calculates total.
6. Process customer payment– EFT, credit card or cash.
7. (i) If EFT then swipe card, have customer enter their PIN and wait for approval.
(ii) If credit card then wait for approval receipt and collect customer signature.
Check signature matches signature on credit card.
(iii) If cash payment then enter amount tendered and hand change, if any, to
customer.
8. Hand receipt (tax invoice) to customer.
Generate Purchase Orders – this transaction creates and submits purchase orders to
each supplier electronically.
1. User initiates transaction on a daily basis.
2. System queries database for low stock products. Query returns number of each
product to order sorted by supplier.
3. Review each product and confirm order.
4. System generates and submits purchase orders to suppliers via either email or fax.
Receive Delivery – this transaction processes each order when it arrives at the store.
1. Manually check actual products delivered match delivery docket.
2. Enter purchase order number from delivery docket. System retrieves and displays
purchase order.
3. If invoice and delivery docket products match then enter date received.
4. System adds number of each item to current stock level of each product.
Enter Invoice – this transaction processes each invoice received from suppliers.
Invoices arrive by mail or fax and are often batched entered on a weekly basis.
1. Enter purchase order number from invoice.
2. System retrieves and displays purchase order.
3. If invoice details match purchase order details and products received then enter
invoice number and mark for payment.
4. Details, including prices, that do not match require manual override/correction in
consultation with supplier.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 389

Pay Suppliers – this transaction produces payments for each supplier based on those
invoices that are due for payment. From the user’s perspective this is a batch process
performed at the start of each month.
1. User initiates transaction at start of each month.
2. System retrieves and displays summary of remittance advice notices for payments
due to each supplier. Each remittance advice includes invoice numbers and invoice
totals, together with payment total.
3. User confirms each supplier payment.
4. System generates remittance advice notices that include printed cheques.
GROUP TASK Discussion
Identify the participants and the tasks they complete during each of the
above transactions.

Based on the above four transaction we create a lower level DFD to describe the flow
of data within the above system (refer Fig 4.13 DFD). On this DFD the store database
is included twice – this is simply to improve readability. Clearly other transactions
will also occur in most real world POS systems.

UPC Products
Product Description,
Sell Price, Stock Level
Sales
Payment
UPC, Details
Generate Low Stock Reduced Stock Level
Purchase Query Receipt Customers
Orders
Details
Store
Supplier Details, Database Purchase Order Details,
Low Stock Products, UPC, Current Stock Level
Number to Order

Purchase Purchase Order Number, Date,

Order UPC, Increased Stock Level
Receive
Suppliers Delivery
Delivery Docket
Remittance
Advice
Invoice

Enter
Pay Purchase Order Number,
Invoice Number, Invoice
Suppliers
Payment OK
Payment
Confirmation
Purchase Order Details
Remittance
Advice Details Store
Database
Fig 4.13
DFD for a typical POS system.

GROUP TASK Discussion

Confirm the DFD in Fig 4.13 correctly reflects the context diagram in Fig
4.12 and the transaction descriptions on the previous page.

Information Processes and Technology – The HSC Course

390 Chapter 4

Data/Information
The data entering and used within POS systems is detailed on the context diagram
(Fig 4.12) and DFD (Fig 4.13). These models also show the information output by
POS systems – that is, receipt details, purchase orders and remittance advices.
The data within POS systems is almost always stored within a relational database. For
the system described above tables for products, suppliers and purchase orders would
be required – a possible schema is reproduced below in Fig 4.14. In reality the schema
would be far more complex to meet additional requirements. For instance, currently
no record is maintained of when products were sold and therefore sales trends cannot
be analysed. Also each product is assigned a single supplier and cost price. In reality
many products are available from multiple suppliers at varying prices. Additions and
modifications would also be required if the retailer accepts orders for out of stock
products or high value products that are individually coded. Most POS systems also
maintain records of each sales assistant and the sales they process.
Suppliers Products PurchaseOrders
1
SupplierID 1 UPC m PONumber
Company Description UPC
Address CostPrice NumOrdered
City SellPrice DateReceived
Postcode m StockLevel InvoiceNumber
PhoneNumber ReorderLevel PaymentOK
FaxNumber SupplierID PaymentMade

Fig 4.14
Initial schema for a simple POS system.

GROUP TASK Discussion

For each of the four transactions described and with reference to the
above schema, identify records that are examined, created and/or
modified.

GROUP TASK Discussion

Recommend modifications to the above schema so that records of when
each sale was made are maintained.

Information Technology
The essential information technology for POS
systems includes a database server that runs
DBMS software and includes sufficient storage
to secure and maintain the database. For smaller
retailers backups are made to CDRs, whilst
larger systems include tape drives. One or more
POS terminals are installed which run the client
application that processes sales transactions.
Further personal computers are often present to
perform other transactions. Commonly an
Ethernet LAN is used to connect to the database
server.
Fig 4.15
Touch screen POS terminal.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 391

All hardware, apart from the POS terminals, is common to many other systems.
Therefore we restrict our discussion to the detail of POS terminals. Firstly, the use of
the word “terminal” is somewhat misleading – in fact most current POS terminals are
in fact personal computers that include integrated collection and display devices. In
the past POS terminals were indeed terminals where processing was performed
centrally. Today POS systems are largely client-server systems and hence much of the
processing is performed by the client.
Currently most POS terminals include a standard PC motherboard including Intel
processor, RAM and hard disk. Attached or integrated devices include touch screens,
magnetic stripe readers, barcode scanners, cash drawers, receipt printers and
specialised keyboards.
GROUP TASK Practical Activity
Make a note of POS terminals you observe during the week. Identify the
devices present within these POS terminals and comment on the design of
each POS terminal.
For most POS systems the size and robustness of POS terminals is at least as
important as the technical performance specifications. There is limited space at most
checkouts and POS terminals are used continuously for extended periods. POS
terminals must be better able to withstand spills and other hazards. The small size of
LCD monitors made them popular inclusions in most POS terminals long before their
widespread use for other applications.

Fig 4.16
Example restaurant touch screen user interface.
Ergonomic issues for participants using POS terminals are different compared to the
issues present for those seated at more traditional computer workstations. POS
terminals are commonly used whilst standing for extended periods of time and the
collection devices are different. The tasks performed by POS terminal users often
include a much broader range of movements as they scan products, use touch screens
and interact with customers. Barcode scanners, touch screens and magnetic stripe
readers reduce the likelihood of RSI and other health issues associated with keyboard
Information Processes and Technology – The HSC Course
392 Chapter 4

data entry. The design of user interfaces for touch screen POS applications is quite
different to other user interfaces. For example the screen reproduced in Fig 4.16
includes large coloured buttons and is customised for each restaurant.
GROUP TASK Discussion
Compare and contrast the design of touch screen user interfaces with user
interfaces designed for use with a keyboard and mouse.

LIBRARY LOANS SYSTEMS

Library systems perform a variety of processes in addition to transactions required to
lend resources. Some of these processes include maintaining the catalogue, searching
the catalogue, purchasing, management of library finances, booking of personal
computers and small group rooms, managing and charging for printing, sharing of
collections with other libraries and integration of digital data within traditional
collections. We cannot hope to examine all these areas of library management hence
we restrict our discussion to some of the transactions related to the lending (check-
out) and return (check-in) of library resources. In library terms these processes are
known as “circulation” processes. The desk where resources are checked in and out is
thus known as the circulation desk.
Circulation rules specify how books and other resources move between borrowers and
the library. Examples include the length of time and number of books that maybe
borrowed, the criteria for determining if loans can be extended, which resources can
and cannot be borrowed and procedures for reserving resources that are currently
loaned. These rules are implemented as transactions where all relevant rules must be
observed for a loan to take place.

Consider the following:

The decision table below is used as the basis for approving loans at a particular
library. Blanks on the rules grid below indicate either a tick � or cross � is possible.
Conditions Rules
Borrower is a current library member � �
Borrower has overdue fines owing � �
Borrower has overdue books � �
Borrower has reached their item limit � �
Resource is reserved � �
Resource can be borrowed � �
Actions
Loan approved � � � � � � �
Loan rejected � � � � � � �
An equivalent decision tree is reproduced below:
Resource
Library Overdue Overdue Item limit Resource can be Loan
member fines books reached reserved borrowed approved
Y N N N N Y Y
N N
Y N
Y N
Y N
Y N
N N
Fig 4.17
Example decision table and tree for approving library loans.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 393

GROUP TASK Discussion

Confirm that the above decision table and tree are indeed equivalent. Is
this loan approval process suitable for implementation as a transaction?
Discuss with reference to the ACID properties and the ability of the
transaction to be rolled back or committed.

Check-out Transaction
The DFD reproduced in Fig 4.18 models a possible library check-out transaction. This
DFD includes processes and data flows to implement the rules described in the
decision table (and decision tree) within Fig 4.16. Note that the Check book can be
borrowed process occurs for each book (or other resource) that a particular member
wishes to borrow. The Check member can borrow process occurs once for each
check-out transaction, as does the Approve loan process.
BookDetails,
BookID Reserved,
Check book CanBorrow
Books can be
BookOK, Title borrowed

BookID ItemLimit, ItemsBorrowed

Library
Approve Database
MemberID loan Final Loan Details

Loan Receipt MemberOK

Details MemberDetails,
CurrentMember,
Members OverdueFines,
MemberID Check
member can OverdueBooks
borrow
Fig 4.18
Possible DFD for a library check-out transaction.

GROUP TASK Practical Activity

Based on the above DFD, create a step-by-step description of the logical
order of processing during a library check-out transaction.

Check-in Transaction
When books are returned or checked into the library the only input data required is the
unique identifier for the book. This identifier, say BookID, is sufficient to search the
database for the loan record currently associated with that book. The transaction can
then update this loan record to record the date the book was returned.
Other processes occur after books have been checked-in. For example the library staff
must manually check the condition of returned books as they replace them on the
shelves. Any damaged books are repaired and if the damage is excessive then the
member may well be expected to pay for repair or replacement. The date returned data
is examined when generating overdue notices and overdue fines – these are
commonly batch processes performed every few days.
GROUP TASK Discussion
Propose a list of likely information that could be generated from the data
created during check-out and check-in transactions. Discuss the purpose
of each type of information proposed.

Information Processes and Technology – The HSC Course

394 Chapter 4

Consider the following:

Checkpoint Meto’s Intelligent Library System is based on passive radio frequency

identification (RFID). Each book contains a passive RFID tag which permanently
stores up to 96 bits of data – the data uniquely identifies each book or other resource.
Readers are used to receive the data stored on RFID tags. RFID readers transmit at a
frequency specific to the antenna withn the library’s RFID tags. The passive tags used
within this library system do not contain batteries, rather they are powered by the RF
energy from readers within range. The tag responds by transmitting its stored ID back
to the reader. Therefore in RFID systems, readers and tags are both transmitters and
receivers.
Currently passive RFID tags operate within a range up to approximately 10m
depending on the power of the specific reader. Active RFID systems utilise larger
battery powered RFID tags and are able to transmit their data over much greater
distances – up to 100m is typical. RFID systems are currently replacing many
electromagnetic (EM) security systems in both libraries, retail stores and warehouses.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 395

Fig 4.19
Checkpoint Meto’s Intelligent Library System
Source: Checkpoint Meto (www.checkpointmeto.com.au)

GROUP TASK Discussion

Identify each of the RFID readers present on the above diagram. Discuss
the information processes that use data from each of these RFID readers.

GROUP TASK Research

Research the operation of electromagnetic (EM) security systems. Then
compare and contrast EM systems with RFID systems.

GROUP TASK Discussion

Compare and contrast the above RFID collection system with traditional
manual and barcode collection systems.

Information Processes and Technology – The HSC Course

396 Chapter 4

Data/Information
Consider the check-out and check-in transactions described above. The data described
in the Fig 4.20 data dictionary below is either stored within the library database or is
generated from data within the library database.
Data
Name Description
Type
BookID Integer PK for each library resource
Book Details Record Various attributes including ISBN, Title, Author, Publisher, etc.
Reserved Boolean Is the resource currently reserved.
Can this resource be borrowed (True) or is it for in library use only
CanBorrow Boolean
(False).
BookOK Boolean True if this resource can be borrowed, otherwise false.
MemberID Integer PK and membership number for each library member
Various attributes including member name, address, phone,
MemberDetails Record
membership status and other details.
CurrentMember Boolean True if membership exists and is current, otherwise false.
OverdueFines Currency Dollar amount of each overdue fine for member.
OverdueBooks Text Details of each currently overdue resource.
True if the member is able to borrow more resources, otherwise
MemberOK Boolean
false.
Maximum number of resources a member can borrow
ItemLimit Integer
simultaneously.
ItemsBorrowed Integer The number of resources a member currently has on loan.
Various attributes for a new loan. Includes the MemberID and
FinalLoanDetails Records
BookIDs for each resource, together with the date borrowed.
Loan Receipt List of book titles borrowed together with the date borrowed and
Various
Details due date.
ReturnDate Date Date each resource is returned or checked-in.

Fig 4.20
Data dictionary for data used by the library check-out and check-in transactions.
Back in chapter 2 we examined the design of a relational database for a library
system, refer Fig 2.17 on page 133. The general nature of this schema meets most of
the requirements for our current check-out and check-in transactions.
GROUP TASK Discussion
Identify the data collected, the information produced and the participants
involved in the check-out and check-in transactions.

GROUP TASK Discussion

Identify data used by the check-out transaction (Fig 4.18 DFD) that is
NOT stored within the library database.

GROUP TASK Discussion

Modify the schema developed back on page 133 Fig 2.17 to meet the
needs of the check-out and check-in transactions described above.

Information Technology
The majority of library systems store their data within a relational database managed
by a database server running DBMS software. This library database is accessed by all
users, including library staff, library members and commonly by remote users via the
Internet. Clearly the security of this data is critical to the continued operation of all
libraries, hence fault tolerant hardware and regular and thorough backup processes are
required. The large quantity of data generally requires the use of automated tape
backups where tapes are stored securely off site.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 397

A variety of client machines running dedicated applications that perform specific

tasks are used. Most libraries have a number of machines dedicated to searching the
catalogue (and often the catalogue of other libraries). The library staff perform
circulation and also a large variety of other transactions, hence their machines run
fully featured versions of library applications.
Staff machines require access to printers, collection devices such as RFID readers and
barcode scanners. When new books are purchased it is common for catalogue records
to be obtained electronically from outside sources – such as the Australian National
Bibliographic Database (ANBD). The ANBD not only provides catalogue data to
libraries, it also maintains records of which libraries have copies of particular titles.
Today catalogue data is downloaded from the ANBD via the Internet, therefore an
Internet connection is required for library staff machines.
GROUP TASK Activity
Analyse the hardware and software present in your school’s library system.
Identify examples of real time transactions performed within this system.

HSC style question:

SuperBook is an Internet based service that allows customers to make bookings and
pay for tickets to major music and sporting events.
When visiting the website customers choose an event, view the currently available
seats, choose their desired seats and finally purchase tickets. It is critical that the
displayed available seats are wherever possible one hundred percent correct.
In relation to the SuperBook service:
(a) Identify the required information technology.
(b) Analyse the SuperBook service in terms of maximising data integrity.
(c) Construct a data flow diagram for the SuperBook service that describes the data
movements between customers, processes and the SuperBook database. Your
data flow diagram should include the following processes:
– Choose Event
– Display Available Seats
– Choose Desired Seats
– Purchase Tickets
Suggested Solution
(a) Information technology includes:
• Database server running DBMS server software to access the databases
containing the seating and bookings for each event.
• The server should include redundant mirrored hard disks to ensure fault
tolerance.
• Web server that creates and transmits web pages to each user based on their
selections. The data required to create each page is retrieved from the database
server.
• The web server includes encryption software so that payment details are
secured during transmission and also once stored.
• The customer requires a machine with web browser and Internet connection.

Information Processes and Technology – The HSC Course

398 Chapter 4

(b) To maximise data integrity:

• A transaction should be used that commences when the customer selects a seat
or seats and ends when the payment has been confirmed. This ensures the
same seat cannot be booked by multiple customers because no other
transactions can access the seat unless or until the transaction fails.
Furthermore the final booking is guaranteed to be stored permanently due to
the durability property of transactions.
• It is impractical to lock all records related to all available seats, hence once a
customer selects their seats it is necessary to validate that they indeed remain
available as another customer may have selected the same seat after display of
available seats. Records for selected seats can then be locked.
• Validation of customer details and payment details is also required to ensure
the accuracy or integrity of the data. The payment details are validated during
the payment approval operation. The customer details should be displayed to
the customer so that they can verify they are correct. Some sites require
customers to respond to an automatically generated email – this ensures that at
least the email address is correct.
(c) Chosen Event
Event Details
Choose
Event

Event ID

Display Available Seats SuperBook

Available Seats Available Database
Seats
Customers
Event ID,
Available Seats
Chosen Seat
Choose
Desired
Seats
Event ID, Seat ID,
Customer Details, Event ID,
Customer Details,
Payment Details Seat ID
Payment Details
Purchase
Ticket Details Tickets

Comments
• In an actual trial or HSC examination part (a) and (b) would be worth 2 or 3 marks
each, part (c) would be allocated 4 or 5 marks. Therefore a total of between 8 and
11 marks would be allocated to this question.
• Answers to parts (b) should specifically address the requirement in the question
that available seats are wherever possible one hundred percent correct.
• In part (a) the keyword “identify” requires recognising and naming the information
technology likely to be present, however unless absolutely obvious it is worth
including a brief justification for inclusion of each item.
• In part (c) no data is written to the database until all processes have completed.
This occurs as the transaction is committed. The transaction log and associated
data flows could have been included on the DFD, however given the specifics of
the question this is unlikely to be required to gain full marks.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 399

SET 4B
1. Generally real time transaction processing 7. DFDs produced from a context diagram
requires: should always include:
(A) fast direct access to storage. (A) identical data flows entering and
(B) secure communication channels. leaving the system.
(C) more processing power than batch (B) the same number of processes.
systems. (C) at least one data store.
(D) All of the above. (D) the external entities that are shown on
2. On the Fig 4.7 context diagram (page 382), the context diagram.
which of the following is true? 8. Consider the initial schema for a simple POS
(A) Each external entity is just a source. system shown in Fig 4.14 on page 390.
(B) Each external entity is just a sink. When a purchase order is created that
(C) Each external entity is both a source includes 4 different products, which of the
and a sink. following is always true?
(D) There is one external entity and two (A) 1 record is created in the
processes. PurchaseOrders table.
3. Large buttons are preferred on user (B) 4 records are created in the
interfaces for which device? PurchaseOrders table.
(A) Touch screen (C) 4 records are created in the
(B) LCD screen PurchaseOrders and Products tables.
(C) CRT monitor (D) 4 records are created in the
(D) Printers. PurchaseOrders and Products tables
and 1 record in the Suppliers table.
4. Consider the schema in Fig 4.10 on page
386. In addition to the primary key, which 9. On the DFD in Fig 4.18 on page 393, which
attributes in the availability table are of the following best explains why there are
populated during the reservation transaction two “Check…” processes feeding data to the
described in the text? “Approve Loan” process?
(A) GuestID, StayDate, RoomID (A) There are only two decisions to be
(B) GuestID, StayDate, RoomTypeID made prior to approving a loan.
(C) GuestID, RoomID, RoomTypeID (B) The DFD would become too complex
(D) StayDate, RoomID, RoomTypeID if all required decisions where detailed
on the DFD.
5. UPCs are often printed as barcodes on the
(C) One process executes for each book
packaging of products. The purpose of UPCs
and the other executes once for each
is to:
loan to check the borrower is OK.
(A) identify different products uniquely.
(D) DFDs should not model the intricate
(B) identify individual items uniquely.
details of all processing.
(C) encode the price of each product.
(D) improve product security. 10. Which of the following CANNOT be
controlled when collecting data over the
6. Most current POS terminals can be best
web?
described as:
(A) Speed of data access from server
(A) dumb terminals that perform only
secondary storage devices.
collecting and displaying processes.
(B) Speed and reliability of Internet
(B) personal computers with specialised
connections.
collection and display devices.
(C) The number of concurrent users that
(C) a combination of specialised collection
can be supported.
and display devices.
(D) The isolation transaction property as
(D) intelligent terminals that perform
many users can read and alter the same
minimal processing such as data
data simultaneously.
validation.
11. With reference to the DFD in Fig 4.13 on page 389, construct a lower level DFD to model the
Sales process.
12. Outline reasons why many POS terminals use touch screens in preference to keyboards.
13. Visit a local supermarket, hardware or department store. Identify the information technology and
participants within the TPS.
14. Analyse your school or local library’s check-out transactions. Determine the information
technology and describe the operations performed during a typical check-out transaction.
15. Examine a web-based reservation system for an airline or rental car company. Construct a data
dictionary detailing the data collected. Include a column explaining the purpose of each data item.

Information Processes and Technology – The HSC Course

400 Chapter 4

BATCH TRANSACTION PROCESSING SYSTEMS

Batch transaction processing separates data collection from the actual transaction
processing operations. This allows transaction processing to be delayed, perhaps to a
time when the system is quiet, such as late at night, or perhaps until sufficient data
exists, such as generating bills in batches of 100, for example. In all cases the data
required for many individual transactions of the same type is collected over time and
then at some later time all transactions are processed together without any user
interaction. Common examples of batch processing include clearance of cheques, bulk
generation of bills and payments, and payroll processing. Batch processing is common
when the required data is collected on paper forms.
Commonly the collected data for each individual transaction is added to a transaction
file. This can occur over time as data is collected or it can occur just prior to batch
processing commencing. When the time comes for batch processing to occur the
transaction file is submitted for processing. This file provides all the input data
required to process all transactions. During batch processing each transaction executes
in much the same way as a single real time transaction. Successful transactions are
committed (permanently written) to the master file or database. Transaction problems
and errors result in rollback of the transaction. Details of the problems that caused
rollback are written to an error file (or log). Writing details to a file allows further
transactions within the batch to continue without user interaction. The systems
flowchart in Fig 4.21 describes the typical processes occurring during batch
transaction processing.

Retrieve next
transaction Transaction
record File

Process
transaction

Commit Yes OK to No Rollback and

Master to master commit store details of Error
File file ? problem File

Yes More
transactions
?

No
Fig 4.21
Systems flowchart modelling typical batch transaction processes.

GROUP TASK Discussion

Compare and contrast real time and batch processing systems with regard
to how transaction problems and errors are resolved.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 401

Historically batch processing was the first type of transaction processing. In the early
days of computers all input was via punch cards – this included the actual program
code as well as the data to be processed. Each card was manually punched by an
operator in preparation for input (see Fig 4.22). Completed stacks of punch cards were
physically loaded into the computer and processed sequentially. In these early days
online real time processing of multiple transactions was simply not possible. The
hardware performed a single task at a time and the output was stored sequentially on
magnetic tape. As a consequence problems associated with multiple transactions
accessing the same data simultaneously did not exist – that is, the isolation ACID
transaction property was simply not an
issue. Furthermore the processing
resources were limited and also costly,
therefore batch jobs were scheduled to
maximise the use and efficiency of
precious processing resources.
Today batch processes are generally
performed in parallel with other
processes. As a consequence ACID
properties must be observed during
most batch jobs, including the isolation Fig 4.22
property. Consider the scenario where a Operators using key punch machines to create
number of different organisations in punch cards for batch processing in the 1960s.
various locations are processing transactions that access the same data. For example
the same credit card number may form part of batch transactions in locations in many
different countries. If these transactions happen to overlap then without the ACID
isolation property data integrity problems will result.
GROUP TASK Discussion
In old systems each transaction in a batch is performed sequentially. Are
any of the ACID properties required within such systems? Discuss.

The processing resources of all computer systems are limited, therefore batch
processing even today is scheduled to ensure that each set of batch transactions can
complete in a timely fashion. This means many batch processes are scheduled to
occur during evenings or weekends when real time processing requirements are
lowest. Such scheduling not only ensures CPU processing resources are available, it
also reduces the wait time for transactions as it is less likely that other transactions
will be simultaneously requiring access to the same data.
Batch transactions that are restricted to a single organisation can be processed offline.
This was the normal situation prior to the widespread use of high-speed
communication links between organisations. Consider a company’s bill generation, all
data originates from one single organisation’s database. In this case a static snapshot
copy of the database can be used to generate bills. Any sales that occur during the bill
generation process are not included until the next batch bill generation process occurs.
User interaction with batch processes is restricted to input prior to the commencement
of processing and to deal with problems after batch processing completes.
Furthermore employees rather than customers commonly initiate batch processing. As
a result the design of user interfaces for batch processing is different – they are
designed for rapid entry. Often such screens accept numerically coded input via the
keyboard’s numeric keypad or a barcode scanner. Screen elements designed
specifically for mouse input are avoided and keyboard shortcuts are available.

Information Processes and Technology – The HSC Course

402 Chapter 4

In this section we examine three examples of common batch transaction processing.

We examine the processes occurring to clear presented cheques, the generation of
bills or invoices and credit card transactions where they appear to be real time but are
often processed in batches.
CHEQUE CLEARANCE
Cheques are a paper-based system that prior to the widespread use of computers were
the primary method for exchanging funds over long distances. Since the 1970s credit
cards and EFTPOS have steadily and significantly cut the total value and number of
cheque transactions. In 1994 the daily value of cheque transactions in Australia was
approximately $25 billion, in 2004 the daily value had reduced by four fifths to just
$5 billion. Despite this reduction, cheques are still expected to form a significant
proportion of total financial transactions for the foreseeable future.
A cheque is essentially a promise by a payer (the person or organisation writing the
cheque) to pay the payee (the receiver of the cheque) some amount of money. Cheque
clearance processes commence when a payee deposits a cheque at their local bank
branch. The purpose of cheque clearance processes is to expedite the secure transfer
of funds from the payer’s account to the payee’s account via the banking and
clearance network.
However cheques are promises and promises can and are broken. In the case of
cheques this regularly occurs when the payer has insufficient funds in their account to
cover the value of the cheque. Other problems can also occur, such as forged,
unsigned, illegible, altered, lost and stolen cheques. Cheque clearance processes
include safeguards to identify and deal with such problems.
Prior to the late 1990s cheque clearance processes did not involve any electronic
communication between banks and cheques took 5 or more working days to clear.
Today cheque details are exchanged electronically between banks and most cheques
clear into payee’s accounts in 3 working days. Compared to totally electronic
transfers, such as EFTPOS and direct deposit, 3 working days is an eternity. This
reality is a consequence of the manual processing inherent in a paper-based system.
The value of the cheque must be determined, commonly using scanners and OCR
software or in some cases it is manually entered. Signatures and dates require
verification and the actual paper cheques are physically exchanged between banks.
Until the 1990s all cheques were physically returned to the payer’s branch for
clearance. Today banks operate their own central facilities that perform cheque
clearance processes for many branches.
Financial transaction clearance procedures in Australia are legislated by Government
and controlled by the Australian Payments Clearing Association Ltd. (APCA). All
major banking and financial institutions, including the Reserve Bank, are APCA
members. APCA operates a number of clearance systems, including the Australian
Paper Clearance System (APCS) used primarily for the
clearance of cheques.
Typical steps for cheque clearance in Australia include:
1. Payee Fred receives a cheque in the mail from payer
Freda (who has an account with DEF bank). Fred
deposits the cheque into his account at his local branch
of ABC bank.
2. In some cases the teller immediately passes the cheque
Fig 4.23
through a MICR (Magnetic Ink Character Recognition) MagTek’s Mini-MICR
reader to determine the payer’s BSB (Bank State cheque reader.
Branch) number and account number and the teller
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 403

manually enters the value of the cheque. When this occurs the funds are
immediately credited to Fred’s account as unavailable funds. Usually these funds
immediately begin to accumulate interest in Fred’s account. More commonly the
cheques, together with the deposit slip, are simply filed for later batch processing.
3. During the afternoon all cheques deposited at local branches of ABC bank are
physically transported to a central outwards processing facility operated by ABC
bank. Some smaller banks share such facilities with larger banks.
4. At ABC bank’s outwards processing facility high speed MICR (Magnetic Ink
Character Recognition) readers read payer BSB (Bank State Branch) numbers and
account details from each cheque. Scanners automatically determine the value of
the cheque and the details on deposit slips. Each cheque is encoded with its own
unique ID so it can be traced should it later be dishonoured or stopped. Most banks
also print the cheque value on the
cheque using MICR printers. Payee
accounts are credited with funds if this
has not already occurred at the branch.
Based on the BSB numbers, cheques
are automatically sorted into bundles
destined for different banks together
with the total value of each bundle.
Fig 4.24 shows IBM’s 3890 sorter
which is able to read MICR characters
(Fig 4.25) and sort up to 2400 cheques
per minute. Note that completion of
this batch process provides electronic Fig 4.24
records of all cheques deposited into IBM 3890 high speed cheque sorter includes
payee accounts operated by the bank. MICR reader and optional scanner.
5. Each bundle of cheques is transported
to a central check clearing house operated by APCS. Appointed representatives of
all banks exchange bundles of cheques. In addition the net difference between
exchanged bundles is calculated. For example the
representative from ABC bank may hand the DEF bank
123467890
representative cheques totalling $2.2 million, whilst DEF a BSB
bank hands ABC bank bundles of cheques totalling $2.5 b Amount
million. In this case the net difference of $300,000 is c Domestic
transferred from ABC bank to DEF bank. At this stage d Dash
all cheques are now under the control of the payer’s
Fig 4.25
bank. In our example Freda’s cheque is now in the hands Standard MICR characters.
of her bank – DEF bank.
6. Bundles of cheques are now physically transported to the central inwards
processing facility of each bank – Freda’s cheque goes to DEF bank’s inwards
facility. Currently most facilities are within major cities such as Sydney and
Melbourne. The cheques commence being batch processed. Each cheque again
passes through a MICR reader and scanner. The scanner determines the value of
each cheque whilst the MICR reader determines the account. For each cheque, the
system ensures there are sufficient funds in the payer’s account, verifies the
authenticity of the cheque and debits the value of each cheque from the payer’s
account. Problem cheques are diverted for manual examination. Cheques where
there are insufficient funds or other problems are dishonoured. The ID encoded by
the payee bank is used to determine and inform the payee’s bank of such problems.

Information Processes and Technology – The HSC Course

404 Chapter 4

In the past cheques were sorted into individual branch bundles and physically
transported to branches for final batch processing. Today account details and images
of account holder signatures are available online, therefore verification can now take
place centrally via secure communication links. It is the removal of the need to
physically transport cheques back to their branch of origin that has reduced clearance
times from 5 days to the current 3 days.
GROUP TASK Discussion
Each cheque passes through two distinct batch processes. Identify and
describe the operations performed during these batch transactions.

GROUP TASK Discussion

Identify the information technology, data/information and participants
involved in the cheque clearance process.

GROUP TASK Research

Many countries, including Australia, are analysing systems that digitise
images of entire cheques – in the banking industry this is known as
“cheque truncation”. Some countries have already implemented such
systems. Research advantages of such cheque truncation systems

BILL GENERATION
In many systems the generation of bills or invoices is well suited to batch processing.
When orders for products or records of services provided are already within the
system then no extra data collection is required prior to generating invoices. No user
interaction is needed and multiple invoices are usually generated at the same time.
Often bills are generated during times when the resources of the system are not being
used – commonly during the night. Consider telephone, electricity, gas, rates and
other regular household bills. The data exists within the organisation’s database and
therefore batch processing can be used to generate bills.
Even small businesses that process small numbers of orders each day use batch
processing. The orders are entered as they are received throughout the day and then in
the afternoon all the day’s invoices are printed as a batch job. The orders are packed
manually using details from the printed invoices. Each order is then dispatched
together with the invoice. The invoicing database schema we produced in chapter 2
when describing normalisation is typical of such a system (refer Fig 2.70 on page
185). This database would be queried to return all invoice details for the current day.
This query is then used as the record source for a report that generates and prints the
day’s invoices.
GROUP TASK Discussion
Analyse the Invoicing database created in chapter 2 to determine the data
and processes required to generate invoices for the current day.

Apart from the relatively static product details and prices the data required to generate
each invoice is largely independent of the data on all other invoices. This data
independence means that invoices can be generated in any desired order and more
significantly multiple invoices can be generated simultaneously. This characteristic is
particularly significant for large systems that generate many thousands of bills. To
generate say monthly telephone bills requires reading each customer’s address details
and records of all the calls made within the billing period. The batch process does not
need to access, update or create data in any other system. Also during processing no
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 405

data is updated or created within the telephone company’s database that is accessed
during the generation of any other customer’s phone bill. This processing
independence means parallel processing can be used to drastically reduce the total
processing time required.

Consider the following:

For large batch systems, where many thousands of bills are generated in a single job,
it is common to make a snapshot copy of the live data. This snapshot is an offline
copy of the actual data as it was at the end of the billing period – maybe the end of a
year, quarter, month, week or even
the end of a day. The online version
of the database continues to operate
Split Offline Snapshot
Customers Batch bill
Customer
without its performance being A-L data generation
Bills
degraded by the batch processes,
and the batch processes are not
interrupted by the online processes. Batch bill
Customers Customer
Because of the independence of the M-Z data generation
Bills
data and processing the snapshot
copy can be split into different parts Fig 4.26
that are physically stored and batch Parallel batch processing using a split offline
processed in parallel on different snapshot copy of the online database.
storage devices and using different
CPUs (refer Fig 4.26). This distributed processing strategy reduces processing time
significantly – if a batch job takes two hours to complete on one machine then it will
take approximately half this time if two machines are used. Such systems use high-
speed digital printers that link with automatic folding and envelope insertion devices.
The time taken to retrieve records from secondary storage is a significant limiting
factor in terms of improving the performance of large systems. Splitting and
physically storing parts of the database on different storage devices is one technique
that improves performance. Another significant technique used by large batch systems
is sequentially accessing data. This involves accessing data in the order in which it is
physically stored on the disk. In the past, data was physically stored sequentially on
tape and hence it was either impossible or extremely inefficient to read the data
randomly – rewinding and fast-forwarding the tape takes time. Sequential access was
a necessity rather than a choice if jobs were to complete in a reasonable amount of
time – even for small batch jobs. Today the read/write heads within hard disks are
able to quickly jump directly to required records, however this still takes time.
Furthermore hard disks read data in complete sectors and commonly multiple adjacent
sectors are also read. This data is stored in the drive controller’s cache. If random
access is used then much of the data that is physically retrieved is not actually
processed. If sequential access is used then all data retrieved is subsequently
processed and the movement of read/write heads is minimised. Note that significant
performance gains are only possible when the transactions processed are independent
of each other and the data they access is physically stored sequentially. Transaction
processes that use retrieved data as the criteria for searches and that write data require
careful analysis and design if the advantages of sequential access are to be maximised.
For instance, the order in which processes are performed can be significant or it may
be more efficient to remove an operation from a transaction and perform it separately
on all the data. For independent processes, such as those required for bill generation,
the ACID properties can be relaxed somewhat in order to improve performance.

Information Processes and Technology – The HSC Course

406 Chapter 4

GROUP TASK Discussion

Many people now receive bills online via the web, in general these bills are
still the result of batch processing. Why not generate these bills online and
in real time? Discuss.

CREDIT CARD TRANSACTIONS (REAL TIME OR BATCH?)

From the customer’s perspective credit card transactions appear to be processed in
real time. The customer presents their card to the sales assistant who then swipes the
card through an EFTPOS terminal and has the customer sign a receipt. If the signature
on the card matches the signature on the receipt then, from the customer’s perspective,
the transaction is complete. On the web, customers enter their credit card details into a
secure web page and then within seconds the transaction is approved – it appears the
entire transaction is complete. In reality many credit card transactions are batch
processed during the evening following the sale, whilst some are actually completed
in real time. Note that a credit card transaction is not complete until the funds have
moved into the merchant’s account.
All credit card transactions involve at least four significant parties.
Customer
Customers who are credit card holders, merchants who are
generally retailers, card issuers who manage the customer side of

6 Customer Receipt
credit card transactions and acquirers who manage the merchant
side of transactions. Most acquirers and issuers are banks who share

1 Permission
the expense of operating the network and technology between
issuers and acquirers predominantly via the MasterCard and Visa
systems. Let us consider the general sequence of events that occurs
to process credit card transactions (refer to Fig 4.27):
1 Customer gives merchant permission to access credit in their
Merchant
account to pay for goods or services. For “card present”
transactions handing over the card and signing verify that
2 Transaction Details

permission has been given. For “card not present” transactions,

such as telephone and mail order, the verbal or written order and
8 Funds Transfer

credit card details are sufficient verification of permission.

5 Authorised

2 Merchant creates and transmits transaction details manually or

electronically to their acquirer. This can occur via an EFTPOS
terminal, manually by written voucher or over the Internet via a
payment gateway. For larger value manual transactions approval
(steps 3, 4 and 5) are performed over the phone. Acquirer
3 The acquirer receives the transaction and determines the card
issuer. The transaction details are then forwarded electronically
3 Transaction Details

to the card issuer.

7 Funds Transfer

4 The card issuer checks the customer has sufficient credit

4 Authorised

remaining to cover the transaction and reserves these funds. An

authorisation code is sent back to the acquirer.
5 The acquirer receives the authorisation code and electronically
forwards it back to the merchant. On EFTPOS terminals the
word “APPROVED” is commonly displayed. For manual
Card Issuer
transactions approval is given over the phone.
6 Merchant receives approval message generates a receipt and Fig 4.27
hands it to customer. If customer is present then they first sign Communication
the merchant’s copy of the receipt and the merchant verifies the during a credit
signature against the signature on the card. The receipt includes a card transaction.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 407

unique number that identifies the transaction within the merchant, acquirer and
issuer’s systems.
7 The card issuer transfers the funds out of their account and forwards the funds to
the acquirer. Often many transactions are batch processed together hence a single
large transfer takes place together with details of individual transactions.
8 The acquirer deposits the value of each of their merchant’s transactions into each
merchant’s account. In most cases this occurs each evening to finalise the days
transactions.

Consider the following:

The above sequence of steps occurs for all credit card transactions however there are
many different systems that perform these steps at different times and perform some
or all of the steps as batch processes. In some cases other organisations are involved
that relay data between merchants and acquirers or to perform processing on behalf of
merchants.
Let us consider some typical examples and highlight when real time processing and
when batch processing is used:
• Retail EFTPOS terminals supplied and connected directly to a particular bank use a
combination of real time and batch processing for credit card transactions. When a
customer’s credit card is swiped the terminal communicates with the bank
(acquirer) via a telephone line to authorise the transaction in real time. The bank
transmits a reference retrieval number (RRN) back to the EFTPOS terminal and the
terminal displays “APPROVED PLEASE SIGN”. The customer signs the receipt
and the retailer verifies the signatures on the card and receipt match. If the
signatures do not match then the transaction is reversed – this reversal is another
transaction sent to the acquirer.
At the close of business each day the EFTPOS terminal “settles” with the acquirer
bank. The settlement process transmits details of all transactions to the bank. The
bank then batch processes all transactions during the evening resulting in the funds
(less any bank charges) being deposited into the retailer’s merchant account.
• Most retailers now use EFTPOS terminals for their credit card transactions as
described above, however manual systems are still available as a fall back should
the EFTPOS terminal or link to the bank fail. Using a manual system the retailer
manually takes an impression of the customer’s card on a voucher. The voucher is
manually completed by the retailer and then signed by the customer Each retailer
has a floor limit. If the total value of the transaction is above the floor limit then the
retailer telephones the bank for manual authorisation. If authorised the bank reads
out an authorisation number, which is manually written on the voucher. Each
voucher includes the original, which is later submitted to the bank, a copy for the
customer and a copy for the merchant.
At the close of business the retailer completes a merchant voucher that includes the
total number and value of all vouchers. The merchant summary together with the
original of all vouchers is then deposited at the retailer’s local bank branch
(acquirer). The vouchers are batch processed by the bank during the evening.

GROUP TASK Discussion

Why do you think the batch settlement process is performed, why not
simply complete each transaction in real time as it occurs? Discuss.

Information Processes and Technology – The HSC Course

408 Chapter 4

• Some retailers are authorised by their bank to accept mail, phone or fax credit card
orders. These are known as MOTO (mail order telephone order) merchant
accounts. Banks scrutinise retailers more thoroughly to verify that they are
trustworthy and honest before MOTO merchant accounts are approved. Once
approved the retailer is able to initiate credit card transactions without the card
actually being present – just the credit card number and expiry date is required.
The details of the transactions are manually entered into the EFTPOS terminal or
can be manually written onto a voucher. As less information is available to verify
each transaction the retailer must agree to accept a higher level of risk should
transactions be disputed. The transactions are processed similarly to above,
however banks often charge higher rates compared to “card present” transactions.
• Internet credit card transactions for large volume businesses are usually processed
in real time. Commonly the merchant’s website collects details of the purchase,
such as products and prices. The website then directs customers to a payment
gateway which completes the actual financial transaction such that the funds are
moved immediately from the customer’s account into the merchant’s account. This
transfer involves both the authorisation and funds transfer steps occurring
simultaneously and immediately.
GROUP TASK Discussion
Banks view Internet credit card transactions as high risk. Propose reasons
why this is the case? Does real time processing reduce the risk? Discuss.

• Other Internet credit card transactions, particularly for smaller businesses, are
actually processed manually using the retailer’s existing EFTPOS terminal and
MOTO merchant account. The credit card details and the details of the purchase
are transmitted securely to the merchant without any interaction with banks. The
merchant then initiates the transaction manually via their EFTPOS terminal. Such
transactions are settled, along with any in-store purchases, during the evening
using batch processes.
• Businesses that charge customers on a regular basis use batch processing. In this
case the business creates a file containing the details of multiple transactions. This
file is uploaded to the merchant’s acquirer bank where it is batch processed during
the evening. The business must hold an authority from each customer to perform
each transaction. Such batch systems are used for purchases that require regular
payments, for example topping up toll card accounts, making loan repayments and
for payment of telephone, electricity, rates and other regular bills.
GROUP TASK Discussion
The above system does not use real time processing at all. The
transactions are entirely batch processed. Discuss advantages for the
customers, merchants, acquirers and issuers.

• Private companies now provide EFTPOS services and dedicated terminals to

retailers. Often these EFTPOS terminals connect via the Internet and are operated
by or connected to Internet payment gateways. Transactions performed on such
terminals are processed in real time in much the same way as web transactions
performed via payment gateways.
• Companies such as PayPal offer credit card processing services that do not require
merchants to have their own merchant account with an acquirer. Rather the
company uses their own merchant account and acquirer to process transactions on

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 409

behalf of other merchants. These systems generally cost more per transaction and
hence are used by individuals and businesses that process credit card transactions
infrequently.
GROUP TASK Discussion
Analyse each of the above systems and identify where real time processing
is being used and where batch processing is being used. Discuss the
appropriateness of each type of processing for the given system.

HSC style question:

BigBizzCorp is a medium sized business which uses a traditional batch payroll system
to produce weekly payslips for each of its 200 employees who work in one of 10
departments.
Each day when the employees come into work, they ‘clock on’ by locating their
employee time card and punching into a special clocking system, which prints the
current time on their time card in today’s position. At the end of the day, the
employee punches their time card again to allow it to print the time they have just
finished for the day.
At the end of each week, the paymaster collects these 200 time cards, and enters the
start and end times for each day for each of the employees into the Payroll system.
When the weekly payroll is run, a single payslip is produced for each employee
showing their hours worked for this week together with their pay, taxation and
superannuation details. An overall summary of the weekly payroll is also produced for
use by management in their budgetary processes.
(a) The data entry screen for entering each employee’s start and end times into the
batch payroll system is reproduced below:

Weekly Payroll for the week ending ../../..

Department: --
Employee number: ------ Employee Name: --------------
--------------
Start Time End Time Total Hours

Monday: --:-- --:-- --:--

Tuesday: --:-- --:-- --:--
Wednesday: --:-- --:-- --:--
Thursday: --:-- --:-- --:--
Friday: --:-- --:-- --:--
Total Hours for the week: --:--

Done Next

Information Processes and Technology – The HSC Course

410 Chapter 4

(i) Identify fields on the above screen where data would be entered directly by
the paymaster. Explain how the remaining fields would be populated during
data entry.
(ii) Propose suitable validation processes that could be performed on the data
entered through this screen. Justify your responses.
(b) The systems flowchart originally created during the development of the above
batch payroll system is reproduced below. The flowchart diagrammatically
represents the steps performed by BigBizzCorp’s batch payroll processing
system.
Hours
worked

Transaction
Sort
Error listing file

Process A

Sorted
Employee Update
Transactions
Master File Employee Master
(by Employee
File
number)

New Employee Weekly Payroll

Employee Payslips summary report
Master file

(i) Explain the processing likely to be occurring within Process A. Refer to the
output produced, including the error listing, as part of your response.
(ii) Describe the method of data access being used each time a file is read from
or written to within the above system.
(c) Analyse the strengths and weaknesses of the current batch system and assess the
effects of altering this system to a real-time system.
Suggested Solution
(a) (i) The paymaster enters the week ending date and then enters just the start and
end time for each day for each employee. The department, employee number
and employee name being populated sequentially from the Employee Master
File – the data entry process progressing to the next employee each time the
Next button is selected.
The Total Hours are generated in real time once each pair of start/end times
has been entered. Similarly the Total Hours for the week would be calculated
by summing the Total Hours fields – this field being updated as each day’s
times are entered.
(ii) Fields to be validated include:
• Week Ending – the date must be a valid date, less than today’s date. If
data is entered incorrectly, the management summary report will have the
incorrect date on it.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 411

• Department number – must be a valid department within the company (eg

between 1 – 10). If data is entered incorrectly, the management summary
report will bill the payroll to an inappropriate or non-existent department.
• Employee number - must be a valid employee number within the
company (eg between 1 – 200). If data is entered incorrectly then during
the update process, the pay and hours will be accredited to the incorrect
employee, or there will be no such employee to match to the transaction,
and the update will fail.
• Each day’s start and end times – must be a valid time, during valid
working hours of business and with end time greater than start time. If
data is entered incorrectly, the calculated hours worked on that day will be
incorrect and the pay will not be correctly calculated. A common error is
to transpose the start and end times, and the validation should check for
this.
• Total hours for the week – must calculate to be within a reasonable
expected range (say 0 – 60 hours). Users could be prompted to check or
re-enter times if the total exceeds the expected maximum.
(b) (i) Process A occurs each time a set of start and end times is entered for each
employee record retrieved from the Employee Master File. First validation is
performed on each field. If the transaction is correct then it is written to the
transaction file – the data would likely include the employee number, the day
and a pair of start and end times. If an error is encountered then a line is
printed – depending on the operation of the UI, errors could include unknown
employee numbers, departments or unreasonable daily total hours. Process A
occurs for each transaction, hence all correct transactions are written to the
transaction file with an error listing being printed of all problem transactions.
(ii) All data access within the system is sequential. Process A reads each
employee record from the Employee Master File one after the other. Process
A also writes transactions one after the other so that the transaction file
contains a sequence of records where each record is comprised of an
Employee Number, day, start time and end time. The sort process arranges
these into order based on Employee Number. These sorted records being
written sequentially to the Sorted Transactions file.
The Update Employee Master File process would read an employee record
(from the Employee Master File) and a sequence of transaction records (from
the Sorted Transactions file) that match the employee number. These records
are joined, a payslip is printed and a new record for that employee is written
to the New Employee Master File. This process repeats sequentially for each
employee in the Employee Master File.
(c) Possible strengths include:
• Only a single computer is required.
• Only a single direct user is required.
• Data required only at the end of each week so system is well suited to batch
processing.
• Simple system that reflects manual processes.
• No real time queries required which suits batch processing.
Possible weaknesses include:
• Dedicated data entry person required.
• Time delay to process payroll.
• No real time queries are possible.

Information Processes and Technology – The HSC Course

412 Chapter 4

• Difficult to alter to meet new requirements – i.e. Designed for the specific
payroll task.
Effects of altering to a real-time system:
• The Employee master file would need to be rewritten as a direct access file.
This would allow queries at any time by management as to who has been
absent this week, or how many hours have been worked so far.
• It would be possible to allow employees to enter their own times directly
through an interactive system that creates a transaction each time an employee
logs on to start or end their day. This eliminates the need for a data entry
person with associated costs, possible bias or errors in the data entry.
• Validation can be done instantly at the time of data entry by the employee,
without the need for a clerk to look back through the transactions and correct
them if they are identified as being in error.
Comments
• In part (a) there are various different ways to interpret the operation of the screen.
As this is a batch system, perhaps each employee is displayed one after the other
and the user has no control over this order – meaning only start and end times are
entered directly as in the above answer. Or possibly the employee number is
entered which causes that person’s name to be displayed ready to enter their start
and end times. Or possibly department could be entered so a sequence of
employee’s within that department is presented.
• In part (a) (i) it is likely that the total hours data, which is calculated from the start
and end times, would not be written to the transaction file. This data is calculated
on the screen for use by the data entry user – it performs a data verification role.
• In part (a) (ii) the validation processes described could include checking the
employee or department exists within the Employee Master File.
• On the systems flowchart a printed error listing is created during data entry of start
and end times. Although this is possible, today it is more likely that during data
entry such errors would be displayed on the screen.
• As the master file is updated problems can occur. During batch processing these
problems are generally directed to an error log (usually a file). The systems
flowchart included in the question does not detail how users are informed of errors
that may occur during the Update Employee Master File process.
• In (b) (ii) it is possible to interpret the Process A read from the Employee Master
File as random access as the data entry screen can be interpreted to be looking up
employees one by one based on the user’s employee number inputs. If this is the
case then the system is not enforcing the employee order as would occur if access
were sequential. Furthermore the transactions are sorted prior to further processing.
This implies that Process A does not collect, create and then write the transaction
records in the order required by the Update Employee Master File process.
• In part (c) there are many other possible strengths and weaknesses of batch
systems, and effects of altering to real time processing that could be discussed.
Notice the three parts to the question – strengths of batch processing, weaknesses
of batch processing, and effects of altering to real time. In a Trial or HSC
examination equal marks would likely be allocated to each of these three parts.
• In a Trial or HSC examination Part (a) (i) and (ii) would likely attract 3 marks
each. Parts (b) (i) and (ii) would attract 4 marks each and part (c) would attract a
total of 6 marks. Therefore this question would form a complete Transaction
Processing Systems questions worth a total of 20 marks.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 413

SET 4C
1. In most batch processing systems the 6. Recurring household bills are particularly
transaction file contains: well suited to batch generation because:
(A) the results or changes made to the (A) such systems include sequential
master file after transaction processing. secondary storage devices.
(B) the data required to process (B) the data required to generate the bills
transactions. already exists within the system.
(C) a copy of all data that has been altered (C) large companies have staff dedicated to
or added to the master file. the bill generation process.
(D) details of all transactions that have (D) most households pay such bills using
been successfully committed. direct deposit or credit cards.
2. User interaction with batch processes 7. Which of the following occurs at check
includes: clearance houses operated by the APCS?
(A) preparing and/or collecting data prior to (A) Bundles of cheques are exchanged
batch processing commencing. between banks.
(B) correcting errors after batch processing (B) Cheques are scanned to determine their
has completed. value.
(C) scheduling when batch jobs should be (C) The value of each cheque is withdrawn
performed. from the payer’s account.
(D) All of the above (D) Funds are deposited into each payee
3. The isolation ACID property can be relaxed account.
when transactions are: 8. The four significant parties in all credit card
(A) processed in parallel. transactions are:
(B) processed sequentially. (A) Customers, retailers, banks and Visa or
(C) performed in real time from multiple MasterCard.
sources. (B) Customers, merchants, clearance
(D) batch processed. houses and banks.
4. Which of the following is the most (C) Customers, merchants, acquirers and
significant reason why cheque clearance issuers.
takes considerably longer than EFTPOS or (D) Payment gateways, merchants, banks
credit card clearance? and card companies such as Visa and
(A) MICR readers are slow compared to MasterCard.
magnetic swipe readers. 9. According to banks, which of the following
(B) Signatures must be manually verified at lists credit card transactions in descending
the point of sale. order of risk?
(C) Ensuring sufficient funds are in the (A) Internet, MOTO, Card Present
payer’s account is performed manually. (B) Card Present, MOTO, Internet
(D) Cheque details are collected from paper (C) MOTO, Internet, Card Present
documents at different locations. (D) Card Present, Internet, MOTO
5. During batch processing, errors detected are 10. Which of the following best describes batch
commonly written to a file rather than processing?
displayed on screen. Which of the following (A) Collecting occurs over some time and
is the best reason why this occurs? then many transactions are processed
(A) To permanently record details of all together at a later time.
errors encountered. (B) Transactions are processed soon after
(B) It allows batch processes to occur when the required data has been collected.
nobody is present. (C) Many similar transactions are
(C) So users are freed to complete real time processed in parallel.
processes. (D) Transactions are added to a queue and
(D) To allow processing to continue are processed in the order in which they
without interruption. were received.
11. Recount the steps that occur once a cheque is deposited until the funds can be withdrawn.
12. Construct a diagram to describe the order of processing occurring to complete a typical “card
present” credit card transaction.
13. Sometimes ACID properties can be relaxed during batch processing. Discuss using examples.
14. Compare and contrast the general nature of real time and batch transaction processing.
15. Explain why systems that collect transaction data on paper forms are suited to batch processing.

Information Processes and Technology – The HSC Course

414 Chapter 4

BACKUP AND RECOVERY

Backup is the process of making a copy of data in case the original is lost or damaged.
Recovery is the opposite of the backup process where the backup copy of the data is
restored and placed back into the system.
We introduced backup and recovery in Backup
chapter 2 when discussing techniques for To copy files to a separate
securing data. In this section we shall secondary storage device as a
concentrate on various procedures used to precaution in case the first
perform backups together with advantages device fails or data is lost.
and disadvantages.
Backups provide a snapshot copy of data at particular points in time. Each backup
copy allows the system’s data to be recovered back to the state it was in at the precise
time the backup copy was made. In the event of total system failure, such as a hard
disk crash or a fire that destroys the data completely, it is important to be able to
recover to a point as close as possible to the time the failure occurred.
The most common reason for total system Early failure due
failure is hard disk failure – in particular to manufacturing Useful period Late failure due to
components that move the read/write faults of operation component wear
heads. It is a fact that all hard disks will
Drive Failure Rate

eventually fail. Research (refer Fig 4.28)

indicates that rates of failure are high with
new hard disks, largely due to
manufacturing faults. Failure rates are
significantly lower for disks that are
approximately 1 to 6 years of age. Failure Approx Approx
1 year 5-7 years
rates then rise again as components begin
Total Operating Time
to wear out.
Fig 4.28
There are many other problems that occur Hard disk failure rates over time.
where backup copies made at different
times in the past are invaluable. For instance when a user makes changes to a file and
later wishes to revert to a previous version. Also viruses are often detected after a
period of time has elapsed. In each case having many historical copies of the data
allows the system (or a single file) to be restored to a previous state. To recover from
the broadest range of possible problems requires backup copies to be made regularly
and each backup to be kept for a reasonable amount of time.
GROUP TASK Discussion
Brainstorm lists of possible failures where the most recent backup is most
useful and another list where older backups are more useful.
Just how often backups are made and for how long they are kept is dependent on the
value and nature of the data. The value of data includes the costs associated with
recreating the data, together with the cost of the system being inoperable. Currently
recreating 10MB of data is estimated to cost on average about $50,000. Furthermore it
is estimated that some 43% of businesses that experience a severe or total loss of data
never reopen. Clearly the importance of performing regular backups and ensuring
they can be reliably restored is critical.
GROUP TASK Discussion
We have mentioned RAID storage at various times throughout our work.
Do mirrored RAID solutions remove the need to make backups? Discuss.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 415

FULL AND PARTIAL BACKUPS

There are three different types of backup that are commonly used within most backup
procedures – full backups, incremental backups and differential backups. Both
incremental and differential backups involve making partial backups.
Full Backup
As the name suggests, a full backup is a complete copy of all data within the system.
This can be a complete image of the entire hard disk(s), including the operating
system, program files, configuration settings and of course data. For most transaction
processing systems it is the data that is of particular value – the software and
configuration settings rarely change and are far easier to restore. Therefore most
businesses perform full backups of all their data files on a regular basis.
Full backups are the easiest to restore should failure occur. The full backup is simply
copied back into the operational system. Unfortunately copying all files takes a long
time and requires large amounts of storage; therefore it is often impractical to perform
full backups on a daily basis. Common backup procedures specify that full backups be
made on a weekly basis, usually commencing on Friday afternoons and for large
systems continuing over the weekend.
Most operating systems store an archive bit along with each file. The archive bit is set
to true when changes are made to a file and when a file is first created. When full
backups are made all archive bits are set to false indicating that a backup copy of each
file has just been made.
Incremental Backup
Incremental backups include making partial backups that copy all files that have
changed or been created since the last backup – the last backup may have been a full,
or partial backup. An incremental backup therefore includes only those files where the
archive bit is true. As a consequence performing an incremental backup is
significantly faster and requires significantly less storage compared to a full backup.
After each file has been copied its archive bit is set to false. Therefore if incremental
backups are performed each afternoon then each incremental backup copies only
those files that have been altered or created since the previous afternoon’s backup was
made.
Before incremental backups can be made it is necessary to first perform a full backup.
Commonly faster incremental backups are then made on at least a daily basis. The
significant saving in backup time is counteracted by the extra time required to recover
the data. During a recover the latest full backup is first restored, then each partial
backup is restored in the order in which they were made. Hence files that have
changed since the full backup progressively overwrite the older versions as each
partial backup is restored.
Differential Backup
A differential backup uses partial backups to makes copies of all files that have been
altered or created since the last full backup. If such partial backups are made each day
then each will contain copies of all files within all previous partial backups since the
last full backup was made. To restore to the most recent backup requires first restoring
the full backup and then restoring just the most recent partial backup.
In terms of archive bits, differential backups copy all files where the archive bit is
true, however differential backups do not alter any archive bits. Therefore over time
one would expect more and more archive bits to be true and hence more and more
files are included within subsequent differential backups. The size of the differential
backup continues to grow until the next full backup is completed.
Information Processes and Technology – The HSC Course
416 Chapter 4

TRANSACTION LOGS, MIRRORING AND ROLLBACK

Recall that transaction logs contain historical details of each transaction performed
including details of transactions that are currently being processed. These details can
be used to restore a transaction processing system back to a consistent state at some
precise point in time – completed transactions can be recommitted or rolled back and
incomplete transactions can be continued or rolled back. Let us consider some disaster
situations where transaction logs, mirroring and rollback assist system recovery.
• Imagine the drive controller on a database server has failed during a busy period –
which is when such faults usually occur. At the time of failure numerous
transactions were incomplete. To recover from this disaster requires the server to
be shut down, a new drive controller installed and the server restarted. The
transaction log is then used by the system to automatically rollback all incomplete
transactions, which returns the data to its most recent consistent state.
What about transactions that span multiple servers and systems? In general, most
transaction systems automatically abort (rollback) actions that have not been
committed after a specified period of time – this deals with most issues. However,
most systems include further safeguards to ensure this occurs. If the transaction
was initiated by the server that crashed then it sends each system involved in each
transaction a message detailing the transaction and the specific actions they should
abort. What about actions performed on the crashed server that formed part of a
transaction initiated by another system? In this case the server informs the
initiating system, who then rolls back the complete transaction.
• A database server is attached to two RAID storage devices. The first RAID device
stores the main databases and uses RAID striping to improve data access speeds.
The second is a mirrored RAID device and is used to store the transaction logs for
all databases. Now suppose a disaster occurs which totally destroys the first RAID
device and all but one of the hard disks on the second RAID device. Because the
second RAID device was mirrored, the remaining hard drive will contain a
complete copy of the current transaction log – I’d copy this to a fresh hard drive!
In this case recovery first requires installing new RAID devices, installing software
and then restoring the data from the most recent set of backup media. We now have
consistent data but it is missing all changes made since the last backup. The
solution to this problem is to use the transaction log to roll forward and recommit
all transactions performed since the last backup.

Consider the following

Backup and recovery protects against each of the following:

• Hardware errors and failure.
• Software errors.
• Physical theft or destruction of hardware.
• Unauthorised or unwanted changes to data, due to viruses or hacking, for example.
• Intentional changes to data that for various reasons need to be reversed.
GROUP TASK Discussion
For each of the above dot points, discuss suitable backup techniques that
will protect the data. Consider the use of secure onsite and offsite storage,
full and partial backups, mirroring and the use of transaction logs.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 417

BACKUP MEDIA
Magnetic tape remains the dominant media for backing up data on large systems,
including most transaction processing systems. Other forms of backup media include
hard disks, CDs and DVDs. Compared to magnetic tape, the limited capacity, lower
data transfer speed and higher cost of these alternatives makes them unviable
alternatives for backup of most large systems. Currently online businesses are
emerging where backups can be made over the Internet. Some large organisations
maintain their own dedicated high-speed communication links to remote backup sites.
Magnetic Tape
Magnetic tape is a sequential access media contained
within cassettes or cartridges and is currently the most
convenient and cost effective media for backup of large
quantities of data. Magnetic storage, including tape was
described in some detail back in chapter 2, therefore we
restrict our discussion to their widespread use for backup
purposes. Fig 4.29
A single inexpensive magnetic tape can store the Various types of magnetic
complete contents of virtually any hard disk; currently tape cartridges.
magnetic tapes (and tape drives) are available that can store in excess of 500GB of
data at just a few cents per gigabyte. Most backup systems compress data prior to it
being written to tape, this compression usually doubles the capacity of most tapes – a
500GB tape can actually be used to backup 1TB of system data.
Tape cartridges encase a much larger surface area of storage material than other forms
of removable storage. The ability to backup such large amounts of data using just one
tape far outweighs the disadvantages of sequential access. In any case both backup
and restore procedures are essentially sequential processes. Furthermore tape
cartridges are light, portable and do not contain complex electronics. This makes the
cartridges suitable for long term and offsite storage.
There are two different technologies currently used to store data on magnetic tape,
helical and linear. In the related Preliminary textbook we discussed the detailed
operation of helical and linear tape drives.
Tape libraries, such as the one shown in Fig 4.30,
include multiple tapes and multiple tape drives. A
robotic system moves tapes between the storage racks Tape
and the tape drives. Such systems allow tapes to be storage
automatically rotated according to the details of the racks
organisation’s backup procedures. The tape drives are
just normal single drives whose operation has been Tape
automated. The use of many standard tape drives drives
improves the fault tolerance of the tape library as
complete drives can be replaced without affecting or
even halting backup processes.
Various different size tape library devices are Fig 4.30
available to suit the backup demands of different Qualstar’s TLS-58132 tape library
stores up to 340 terabytes of data.
information systems. Small tape libraries are
available that hold just four tapes and use a single drive; these devices provide
capacities suited to most small businesses. Larger devices hold hundreds or even
thousands of tapes and contain many drives. Large government departments and
organisations link multiple tape library devices together; such systems hold hundreds
of thousands of tapes and many thousands of tape drives.
Information Processes and Technology – The HSC Course
418 Chapter 4

Hard disks
The use of hard disks for backup has recently become popular for smaller systems.
External hard disk devices are available that connect to a computer via high-speed
USB or firewire ports, whilst others connect directly to Ethernet networks. In terms of
cost these alternatives are still significantly more expensive than tape if the equivalent
level of protection is to be achieved – currently tapes costs tens of dollars each whilst
similar capacity hard disks cost hundreds of dollars each. For backup processes many
hard disks are required. Nevertheless for small business and home backup purposes
external hard disks are now a viable alternative. For larger systems the physical size,
weight and mechanical complexity of hard disks is significant when the media must
be transported to secure offsite storage.
Note that mirrored RAID systems use multiple hard disks to store copies of data.
These systems protect data and provide fault tolerance should one of the mirrored
drives fail. Such systems do not protect data against total system failure and are of no
use when historical data is required to rebuild the system to a prior state. Hard disks
used for backup are configured to perform full backups and partial backups such that
the system (or individual files) can be restored to previous states.
GROUP TASK Research
Research the current cost of external hard disks with a similar capacity to
the hard disks within current personal computers.

Optical Media – CD and DVD

For single machines and small businesses optical media is a popular and low cost
backup solution – most computers include CD/DVD read/write drives and rewritable
DVD media is relatively inexpensive. A single layer DVD stores 4.7 gigabytes and a
double layer DVD approximately 8.5 gigabytes of data, this is sufficient for backing
up most hard disks. Even when data compression is used multiple rewritable DVDs
will likely be needed for a full backup, however a single DVD is generally sufficient
for partial backups.
For backups of larger systems DVD media does not have sufficient capacity to make
it a viable alternative to tape. Currently a single magnetic tape cartridge can store
more than 500GB of data, this amount of data requires in excess of 50 DVDs. Large
government and commercial organisations use tape libraries that backup to hundreds
or even thousands of tape cartridges. Clearly using DVDs would require 50 times
more DVD disks than tapes. This is unviable in terms of quantity of disks but also in
terms of physically moving the disks in and out of drives.
Recordable “write once-read many” (WORM) CDs and in
particular DVDs are used for archiving critical data that must
be archived in an unalterable form for long periods of time –
often exceeding 10 years. In most instances the data must be
archived permanently because of governmental or
organisational requirements. In most cases this data will
rarely be read. Such applications include medical, legal and
taxation records that were traditionally stored on paper. For
larger systems optical “jukebox” devices are available. These
devices include multiple optical drives together with Fig 4.31
Pioneer DRM-3000
automatic disk changers. The Pioneer DRM-3000 shown in optical disk jukebox.
Fig 4.31 includes two DVD drives and 6 magazines that each
holds 50 DVD-R disks.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 419

GROUP TASK Research

Research and then compare and contrast the current capacity and cost of
tape cartridges compared to optical disks.

Online Systems
Businesses are beginning to emerge on the Internet that specialise in providing online
backup and recovery for individuals and small businesses. These online systems
totally automate the backup process for users. All data is transferred via the Internet to
a secure remote site. The remote site then manages the secure storage of the data on
behalf of the individual or business. Clearly the remote site must use some form of
secure and permanent storage. When first using an online backup system a full backup
must be made, which is a time consuming process. After the initial backup,
incremental backups are made at regular intervals – in some cases every time a file is
saved. Such systems enable recovery of different historical versions of individual files
as well as recovery of complete systems.
Large organisations that manage large volumes of critical data maintain complete
operational copies of their entire system at remote locations. Such copies include the
hardware, software, communication lines and data. Data from the original site is
continually backed up via online communication lines to the remote site or sites. This
is the ultimate in fault tolerance as a complete system failure, such as a fire or terrorist
attack, can be recovered instantly by simply activating the backup site.
GROUP TASK Research
Research using the Internet current online backup services. Determine the
capabilities and cost of such services.

BACKUP PROCEDURES
The same backup media should not be used continuously to perform backups. Rather
multiple sets of backup media should be purchased and used. The aim is to maintain
many complete backup copies produced at different times such that the system’s data
can be recovered back to a variety of different past states. If only a single set of
backup media is used then failure of the media can spell disaster. Furthermore many
problems, such as viruses, may go undetected for some time. In these cases a backup
copy produced prior to the problem occurring is invaluable.
A definite backup procedure is required that is documented and applied consistently.
Most backup procedures fail as a result of human error. Therefore it is vital that
backup procedures are thoroughly understood and are simple to apply. It is
particularly important for the people who perform the backups to be aware of their
importance – backups can easily become a chore that are easily overlooked. The
procedure should specify which set of media is to be used for each backup and when
and where backup copies should be stored offsite.
Backup procedures should also specify how backup copies are to be verified to ensure
they will actually work in the event of failure. Commonly the backup software
verifies all data on the media as the backup is being made – essentially after writing,
the data is read back into RAM and compared to the original. Specialised backup
software is available that can be configured to enforce the backup procedure –
including verification. However human assistance is still needed to physically change
the backup media and to ensure media is stored offsite as required. It is advisable to
manually perform a test recovery at regular intervals to ensure recovery operates as
expected. Such recovery tests should be performed using a different media drive – it is

Information Processes and Technology – The HSC Course

420 Chapter 4

possible that tapes or other media will not operate correctly in different drives. All
backup copies will be useless if the backup drive itself fails or is destroyed.
GROUP TASK Research
Research backup software included with your operating system and other
examples of specialised backup software. Outline the available features.

A simple, albeit costly, backup strategy would be to make a full backup to new media
at regular intervals such as every afternoon. Such a system is certainly simple to
implement and for some critical or high value data such a strategy may well be an
appropriate solution. However for most systems a less costly solution that reuses the
backup media is generally preferred. There are three commonly used media rotation
schemes; Grandfather, Father Son (GFS), Round Robin and Towers of Hanoi. We
shall discuss examples of each of these schemes. To simplify our discussion we
assume a single tape is sufficient for completing each backup. In reality each backup
may require multiple tapes, DVDs or some other type and quantity of backup media.
Grandfather, Father, Son (GFS)
This is the most commonly used rotation scheme. GFS rotation requires daily or son
tapes, weekly or father tapes and monthly or grandfather tapes. Full or partial backups
are performed each working day to a son tape, except for the last workday. On the last
workday a full backup must be performed to one of the weekly or father tapes. At the
end of the fourth week a full backup is made to one of the monthly or grandfather
tapes. The set of son tapes is reused each week, the set of father tapes is reused each
month and the set of grandfather tapes is reused each year. Usually the monthly or
grandfather tapes are stored offsite and the weekly tapes are stored onsite within a
safe, however this is varied to suit the needs of the individual organisation.
To implement a GFS rotation within an Mon Tues Wed Thurs Fri
organisation that operates 5 days per week Mon Tues Wed Thurs Week1
Mon Tues Wed Thurs Week2
requires four son tapes, three father tapes and
Mon Tues Wed Thurs Week3
thirteen grandfather tapes. Note there are 13 Mon Tues Wed Thurs Month1
four-week periods in a year, not 12. The son Mon Tues Wed Thurs Week1
tapes are labelled Mon, Tues, Wed and Mon Tues Wed Thurs Week2
Thurs. The father tapes are labelled Week 1, Mon Tues Wed Thurs Week3
Week 2, Week 3 and the grandfather tapes Mon Tues Wed Thurs Month2
Mon Tues Wed Thurs Week1
Month 1, Month 2, … Month 13. After Mon Tues Wed Thurs Week2
making an initial full backup the schedule in Mon Tues Wed Thurs Week3
Fig 4.32 is used to determine the tape media Mon Tues Wed Thurs Month3
used for each afternoon’s backup. Fig 4.32
Weekly and monthly backups should always Grandfather, Father, Son media rotation.
be full backups, however the daily or son backups can be full or partial backups. If a
relatively small amount of data is present then full backups can be used throughout.
When full backups are used just one tape is required to restore data from the most
recent backup, or indeed from any backup. If differential daily backups are made then
two tapes are required to restore to the most recent backup – the last weekly full
backup is first restored followed by the most recent daily differential backup. If
incremental daily backups are used then the most recent weekly full backup is
restored followed by restoration of each of the subsequent incremental daily backups.
Using full daily backups simplifies the restore process at the expense of longer
backups. Using differential daily backups results in slightly more complex restore
processes, but reduces the time taken for daily backups significantly – only files
changed since the last full backup are copied. Using incremental daily backups

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 421

complicates restore processes, but requires less time for each daily backup – only files
changed during the day are copied.
GFS rotation means that recovery operations can restore back to any day in the last
week, any week within the last month and any month within the last year. Usually the
final yearly backup is archived permanently so it is also possible to restore back to the
end of a particular year. At first it may seem unlikely that such ancient data would
ever need to be restored. This is probably true of the entire data, however it is not
unusual for particular files from previous months or years to be required.
Notice that the son tapes are used much more often than the father tapes and the father
tapes more often than the grandfather tapes. This means the son tapes will suffer the
most wear and grandfather tapes the least wear. Some backup procedures specify that
tapes be simply replaced at regular intervals, whilst other GFS procedures promote
son tapes to become fathers and fathers to become grandfathers. Such promotion
strategies mean new tapes are introduced as sons where they are used actively for a
period of time. As they age they are promoted to become fathers where they are less
active. Finally the father tapes are promoted to become grandfathers where they go to
an offsite retirement home to relax quietly on a shelf!
There are numerous different ways to implement GFS rotation. The number of sons
can be increased for organisations that operate 7 days a week or so that backups are
performed more than once per day. The frequency of father backups can be increased
or decreased, as can grandfather backups. Indeed some complex schemes increase the
number of generations to include great or even great great grandfather generations.
The detail of the procedure is determined by the needs of the organisation. For some
organisations losing even a days work would be catastrophic, whilst for others this is
an acceptable risk.
Round robin
A round robin rotation reuses all tapes equally. Each tape is
Mon
numbered sequentially, say from 1 to 5 or maybe Mon to Fri.
Each tape is then used in turn. When all tapes have been used
Tues
the cycle simply repeats that is, tape 1 is used after tape 5 and Fri
the cycle continues. Clearly when just five tapes are used and
backups are made daily then it is not possible to restore data
Thurs Wed
to states more than five days old. Clearly each tape added to
the cycle extends the ability to restore back a further day.
Fig 4.33
This simply strategy is only suited to small businesses where Round Robin
restoration of data back to a particular day is a high priority. media rotation.
For instance if 30 tapes are used then it is possible to restore
back to any day in the past month. In reality most organisations that use a round robin
scheme will (or should) also archive backups permanently at regular intervals.
Towers of Hanoi
This is a somewhat complex method of rotation
based on the Towers of Hanoi logical puzzle. In A
the puzzle a series of disks are stacked in size B
order on one of three poles as shown in Fig 4.34. C
The aim is to move all disks to the third pole, D
however you can only move one disk at a time, E
F
and larger disks can not be placed on top of
smaller disks. In Fig 4.34 we have six disks
labelled A to F in ascending size order, however Fig 4.34
The Towers of Hanoi puzzle.
any number of disks is possible. The solution
Information Processes and Technology – The HSC Course
422 Chapter 4

involves moving the smallest disk A every second move, disk B every fourth move,
disk C every eighth move and so on. In our example, disk B is first moved on the 2nd
move, disk C as the 4th move, and disk F cannot be moved until the 32nd move.
So how does this puzzle relate to backups and tape rotation? Each disk represents a
tape and the order in which the disks are moved determines the order in which the
tapes are used. Therefore in our example, tapes are used in the order shown in Fig
4.35 below – this complete sequence repeats continuously every 32 days. Notice that
tapes used less often will contain data from the more distant past whilst those used
more often contain more recent data.
Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7 Day 8
A B A C A B A D
Day 9 Day 10 Day 11 Day 12 Day 13 Day 14 Day 15 Day 16
A B A C A B A E
Day 17 Day 18 Day 19 Day 20 Day 21 Day 22 Day 23 Day 24
A B A C A B A D
Day 25 Day 26 Day 27 Day 28 Day 29 Day 30 Day 31 Day 32
A B A C A B A F
Fig 4.35
Towers of Hanoi rotation sequence with six tapes.
When performed manually any new tape added to the system becomes the new tape A
and all other tapes move up – A becomes B, B becomes C, and so on. Therefore over
time each tape will eventually be used the same number of times. Furthermore offsite
storage can be specified for particular tapes. For example in our example tape E and F
could be stored offsite in the knowledge that they will only be required every 32 days.

GROUP TASK Discussion

Analyse the above Towers of Hanoi rotation sequence to determine the
historical data that can be restored at different points within the sequence.

Specialised backup software that operates in conjunction with tape libraries usually
supports and recommends the Towers of Hanoi rotation scheme. A magazine is
loaded with sufficient tapes for a complete Towers of Hanoi rotation – 6 tapes in our
example above. The software performs the daily tape rotation automatically by
loading and backing up to the correct tape in the sequence. At the end of the sequence
the complete magazine is placed into secure storage and a new (or recycled) magazine
is loaded into the library. The backup software is also used during recovery
operations. Such backup and restore applications track and are able to display the
different versions of data or specific files that are available to be restored. This means
it is not necessary for staff to understand the complexity of the system to restore data
efficiently. Some advanced systems provide network access to tape libraries for end
users. Such systems allow users to restore historical versions of their own files from
the tape library as required.

HSC style question:

Big Bad Bikes (BBB) imports bikes from overseas suppliers and sells them to the
general public. BBB sells mountain bikes, road bikes, BMX bikes and bicycle
clothing. BBB has a transaction processing system (TPS) to process their sales,
generate purchase orders, supplier payments and produce stock level reports.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 423

Each stock item in the store has a barcode. When new stock arrives from suppliers the
barcode is scanned to update the stock inventory database. A point-of-sale (POS)
terminal is used to record all sales and produce customer receipts.
(a) Represent the TPS of Big Bad Bikes using a context diagram.
(b) Propose and describe a suitable backup procedure which may be employed by
Big Bad Bikes.
Suggested Solution Sale details, Purchase order,
(a) Customer Payment details Supplier Payment Details
Customers BBB’s
Sales Receipt Details Delivery Docket, Suppliers
TPS Invoice

Barcode

Products
(Bikes and
clothing)

(b) Backup is the process of making a copy of files used by the system in case the
original is lost or damaged. There are a number of possible backup procedures
that BBB could utilise, however as the data within the system is not enormous a
Grandfather, Father, Son tape rotation scheme, full backups and offsite storage is
a suitable procedure. To verify backup copies are usable and to simplify recovery
in the event of failure it is recommended that two identical tape drives be
purchased – one installed on the BBB server and another on the owners home
computer.
• As the total data within the TPS is likely to fit on a single tape a total of 20
tapes are required. The first four tapes are the sons and are labelled Mon, Tues,
Wed and Thurs. The next 3 tapes are the fathers and are labelled Week 1,
Week 2 and Week 3. The remaining 13 tapes are the grandfathers and are
labelled Month 1 through to Month 13.
• Each afternoon, except on Friday, a full backup is made of the TPS data to the
corresponding son tape. These tapes are stored on a shelf within the office.
• On the first Friday a full backup is made to the Week 1 father tape and then on
the next two Fridays backups are made to the Week 2 and then Week 3 father
tapes. These weekly father tapes are stored within the safe at the shop.
• Every fourth Friday backups are made to the grandfather month tapes in
sequence. That is, the Month 1 tape first, then the Month 2 tape and so on.
After each grandfather backup is made the owner of BBB takes the tape home,
verifies the tape is readable using their home computer and stores the tape
securely until required the next year.
• At the end of each year a further backup is made onto a fresh tape. This
backup is placed into permanent storage – perhaps in a safe deposit box at
BBB’s bank.
Comments
• In a Trial or HSC exam part (a) would likely attract 3 marks and part (b) 4 marks.
• It is not necessary to include employees of BBB on the context diagram as they are
part of the system – they perform the system’s information processes.
• There are numerous other possible and suitable backup procedures that could have
been proposed and discussed.
Information Processes and Technology – The HSC Course
424 Chapter 4

SET 4D
1. Partial backups are included in: 7. Hard disks are least likely to fail:
(A) Full and incremental backups. (A) when new.
(B) Full and differential backups. (B) during mid-life.
(C) Differential and incremental backups. (C) late in life.
(D) Full, differential and incremental (D) during the first 5-7 years.
backups. 8. A GFS rotation has been used for a full year.
2. Which of the following is always true if 10 This rotation uses 4 daily, 3 weekly and 13
media sets are used and backups are made monthly tape sets. Backups are made each
daily? afternoon. To what points in time can data
(A) Data can only be recovered to a state be restored?
within the past 10 days. (A) Any day in the last month and the end
(B) Data can only be restored to exactly 10 of each month in the last year.
points in time. (B) Any day in the last week, any month in
(C) It is possible to recover to more than 10 the last year and any day in the
points in time. previous month.
(D) More than one set of backup media will (C) Any day in the last week, the end of
be needed to restore data. any week in the last month and the end
3. Which of the following lists is in ascending of any month in the last year.
order by storage capacity? (D) Any day in the last year.
(A) CD, DVD, Tape cartridge. 9. A round robin rotation is used with 30 sets
(B) CD, Tape cartridge, DVD. of backup media. Backups are made each
(C) Tape cartridge, DVD, CD. weekday at midday and again at 6pm. Which
(D) DVD, Tape cartridge, CD. of the following is true?
4. When differential backups are made, which (A) Data can be recovered to any day in the
of the following occurs? past 30 weekdays.
(A) All archive bits are set to true. (B) Data can be recovered if it was created
(B) All archive bits are set to false. more than 15 weekdays ago and has not
(C) Archive bits for changed or new files been altered since.
are set to false. (C) Data can be recovered if it was deleted
(D) No archive bits are altered. from the system more than 15
weekdays ago.
5. Currently the preferred backup media for (D) Data that was altered more than 15
large systems is: days ago cannot be recovered.
(A) Magnetic tape.
10. The Towers of Hanoi rotation scheme
(B) Rewritable optical disks.
described in Fig 4.35 has been used for
(C) External hard disks.
many months. Just prior to the day 25
(D) Mirrored RAID.
backup, how old are each of the 6 backups?
6. It is common practice to purchase two (A) A is 1 day old, B is 2 days old, C is 4
identical tape drives and store one offsite. days old, D is 8 days old, E is 16 days
Why is this? old and F is 32 days old.
(A) If the original is damaged or destroyed (B) A is 1 day old, B is 2 days old, C is 3
along with the data then data can still days old, D is 4 days old, E is 5 days
be recovered. old and F is 6 days old.
(B) So backup copies can be verified on (C) A is 2 days old, B is 4 days old, C is 8
another drive. days old, D is 16 days old, E is 32 days
(C) If (or when) the original tape drive fails old and F is 64 days old.
backups can continue without the need (D) A is 2 days old, B is 3 days old, C is 5
to urgently obtain a new tape drive. days old, D is 1 day old, E is 9 days old
(D) All of the above. and F is 25 days old.
11. Define the terms “backup” and “recovery”.
12. List and briefly describe a variety of issues and faults that can be resolved when backup copies of
data are available.
13. Explain the role of archive bits when performing full, incremental and differential backups.
14. Assess the merits of secure onsite storage and offsite storage of backups.
15. Explain each of the following in relation to backup and recovery.
(a) Transaction log (b) Mirroring (c) Documenting procedures

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 425

COLLECTING IN TRANSACTION PROCESSING SYSTEMS

In this section we consider the information technology used to collect data into
transaction processing systems. We examine the operation and use of specialised
collection hardware. We then discuss the design of data collection forms including
paper, online and web-based forms.
COLLECTION HARDWARE
There are numerous different collection hardware devices used within transaction
processing systems. Examples include keyboards, mice, image scanners, touch
screens, RFID readers, Automatic Teller Machines and fingerprint scanners. Earlier in
this chapter we described RFID readers and tags (refer page 394) within a library’s
transaction processing system. In chapter 3 we examined ATMs in some detail – refer
page 294-295. In this section we restrict our discussion to three specialised collection
devices namely, MICR readers, barcode readers and magnetic stripe readers.
GROUP TASK Review
Review the sections on RFID readers (page 394) and also ATMs (page
294-295). Describe their role in transaction processing systems.

MICR for reading cheques

Magnetic ink character recognition (MICR) has been used on cheques since the 1950s
and it still remains the primary method for high-speed collection of cheque data. The
ink (or toner) used to print MICR characters contains ferromagnetic material – a
material that can easily be magnetised. It is the ferromagnetic material within the ink
or toner that is used by the readers. This means customers or bank employees can
write over the MICR characters without greatly affecting the ability of the MICR line
to be successfully read.
There are strict standards that specify the precise design of MICR fonts, paper
characteristics and positioning of MICR line data on cheques. An agreed standard is
critical as cheques are regularly read by banks and clearance facilities across the
world. Furthermore cheques pass through many MICR readers during clearance
processes. Current standards require that each cheque can be read at least 30 times. In
Australia the Australian Payments Clearance Association (APCA) produces and
maintains the MICR specifications for cheques.
Each cheque is encoded with a MICR line positioned across the bottom of the cheque.
The MICR line includes fields for BSB, account number and cheque number. The
value of the cheque can also be included – for large organisations the cheque value
can be printed as the cheque is created, whilst for other cheques the value can be
added to the MICR line during outwards processing. In Australia the MICR E-13B
font is used, whilst many European countries use the CMC-7 font. The majority of
MICR readers include the ability to read MICR lines in either of these fonts. Both
these fonts include the 10 digits and 4 additional symbols for BSB, amount, domestic
and dash – examples of the E-13B character set was shown back in Fig 4.25.
There are two common technologies within MICR readers, waveform and matrix
readers. Waveform MICR readers first magnetise the entire MICR line by passing it
over a write head. The cheque then moves past a magnetic read head. The read head
converts the magnetic data into a very small electrical signal. Each MICR character is
designed to be significantly different, therefore the electrical signal for each character
has a unique signature waveform. The reader detects each waveform and converts it
into the corresponding digital code for that character. Waveform technology is used
within smaller low speed MICR readers. Commonly cheques are fed into these
readers by hand or via a small hopper.
Information Processes and Technology – The HSC Course
426 Chapter 4

Matrix reader technology is used within most high-speed MICR readers, including the
IBM 3890 and Unisys DP1800 series of reader/sorters. This technology has been in
use since the early 1960s and remains the dominant technology used within most
bank’s readers and sorters. Matrix readers are able to read and sort up to 2400 cheques
per minute. Each MICR line is magnetised such that each character is split into many
vertical slices. The read head includes 30 mini-heads

4
positioned at right angles to the slices. Therefore each
character is split into a matrix of magnetic cells – each slice
split into 30 cells. As each cheque’s MICR line passes
through the reader, each of the 30 mini-heads simply
determines whether each cell in each slice is magnetised.
The result being a mini bitmap of each character (see Fig
4.36). This bitmap is then converted to its corresponding Magnetised
cells
digital code. The data is transmitted back to the system for Fig 4.36
further processing and storage. The BSB data is commonly Matrix MICR readers read
each character as a matrix
used as the basis for sorting cheques in preparation for of magnetised cells.
transport to particular banks or branches.
GROUP TASK Investigation
Examine the MICR line on a cheque. Identify the fields present within the
MICR line and the symbols used to separate these fields.

Barcode Readers
Barcode readers or scanners operate by reflecting light off the barcode image; light
reflects well off white and not very well off black. This is the basic principle
underlying the operation of all types of scanners. A sensor is used to detect the
amount of reflected light; so to read a barcode we can either progressively move the
light beam from left to right across the barcode or use a strip of light in conjunction
with a row of light sensors. Each of these
techniques are used for different designs of
barcode scanner; those based on LED, laser
and CCD technologies dominate the market,
Fig 4.37 shows an example of each. Most
barcode scanners incorporate a decoder to
organise the data into a character
representation that mimics that produced by
the keyboard. This means most barcode
readers can be installed between the keyboard
and the computer without the need for
dedicated interface software.
Barcode wands use a single light emitting
*9350(6440!
diode (LED) to illuminate a small spot on the
barcode; the reflected light from the LED is Fig 4.37
measured using a single photocell. As the Clockwise from top-left: LED wand,
multi-directional laser and CCD based
wand is steadily moved across the barcode, barcode scanners.
areas of high and low reflection change the
state of the photocell. The photocell absorbs photons (a component of light); as the
intensity of photons absorbed increases so too does the current flowing through the
photocell; large currents indicating white and smaller currents indicating black. This
electrical current is transformed by an analog to digital converter (ADC) to produce a
series of digital ones and zeros. The same LED technology is used for slot readers,
where the barcode on a card is read by swiping the card through the reader.
Information Processes and Technology – The HSC Course
 Option 1: Transaction Processing Systems 427

Lasers are high intensity beams of light; as such they can be directed very precisely.
Laser barcode readers can therefore operate at greater distances from the barcode than
other technologies; commonly up to about 30cm away. The reflected light from the
laser is detected by the photocell using the same technique as LED scanners. There is
no need to manually sweep across the barcode as the laser beam is moved using an
electronically controlled mirror. Basic models continually sweep back and forth
across a single path, whilst more advanced models perform multiple rotating sweeps
that trace out a ‘star like’ pattern. These advanced models are much more effective as
the user need not hold the scanner parallel to the barcode; rather the scanner rotates
the scan line until a positive read is collected. Supermarkets often use this type of
barcode scanner mounted within the counter top.
Barcode scanners based on charge coupled Original image or barcode
devices (CCDs) contain a row of photocells
built into a single microchip. CCD technology is
used within many image collection devices
including; CCD barcode scanners, digital still Lamp
and video cameras, handheld image scanners, (or row of LEDs) Mirror

and also flatbed scanners. For both barcode and

image scanners a single row CCD is used. The
light source for these scanners is typically a Lens
single row of LEDs; the light being reflected off
the image back to a mirror. The mirror reflects CCD ADC Digital
output
the light onto a lens that focuses the image at the
CCD. Each photocell in the CCD transforms the Fig 4.38
light into different levels of electrical current The components and light path typical
which are converted into bits, using a similar of most CCD scanner designs.
technique to that used in LED and laser barcode scanners. CCDs in image scanners
differ slightly; they convert the electrical current from each photocell into a binary
number, normally between 0 and 255, using a more complex analog to digital
converter (ADC).
GROUP TASK Research
Research the specifications and cost of different barcode readers. Classify
each barcode reader according to its suitability for use within retail
transaction processing systems.

Magnetic Stripe Readers

Today almost all plastic cards contain magentic
stripes. This includes credit cards, ATM cards, 2.8mm Track 1 (210bpi)
library cards, ID cards, frequent flyer cards, store
Track 2 (75bpi)
charge cards and even some temporary and limited 0.6mm
Track 3 (210bpi)
use paper cards such as train and bus tickets. Apart
from some rare exceptions the magnetic stripes on Fig 4.39
these cards include three parallel tracks. Each track Typical layout and density of
tracks on magnetic stripes.
is approximately 0.11” (2.8mm) wide and 0.025”
(0.6mm) apart (refer Fig 4.39). Tracks one and three contain data written at a density
of 210 bits per inch, whilst track two is written with a density of just 75 bits per inch.
Most ATM and credit cards contain data on tracks 1 and 2, however most readers,
including those within EFTPOS terminals and ATMs are only capable of reading
track 2. Track 2 contains the primary account number and other details unique to the
card issuer – many ATM cards contain an encrypted version of the PIN. The PIN
cannot be decrypted without a decryption key from the card issuer’s bank. Track 1
Information Processes and Technology – The HSC Course
428 Chapter 4

contains essentially the same data as track 2 plus the card holders name, so if your
name ever appears on an offline terminal then the stripe reader must be reading track
1 rather than track 2. Track 3 was originally intended to contain rewritable data such
as details of offline transactions. As all EFTPOS and ATMs are online devices track 3
is rarely used.
Track 1 is encoded using a 6-bit subset of Data Parity Char- Function
ASCII and is able to store 79 alphanumeric Bits Bit acter
characters. Track 2 is encoded using 4-bit 0000 1 0 Data
BCD (Binary Coded Decimal) and is able to 0001 0 1 Data
store 40 characters, track 3 also uses BCD 0010 0 2 Data
encoding and can store up to 107 characters. 0011 1 3 Data
The 4-bit BCD character set only includes 0100 0 4 Data
16 characters - the 10 digits and 6 control 0101 1 5 Data
characters as shown in Fig 4.40. All 0110 1 6 Data
characters are followed by an odd parity bit, 0111 0 7 Data
therefore on track 1 a total of 7-bits are used 1000 0 8 Data
1001 1 9 Data
per character, whilst on tracks 2 and 3 just 5
1010 1 : Control
bits are used per character. The data on each
1011 0 ; Start Sentinel
track is followed by a longitudinal
1100 1 < Control
redundancy check (LRC) character. LRCs 1101 0 = Field Separator
calculate an odd parity bit for each 1110 0 > Control
corresponding bit (or column) in each 1111 1 ? End Sentinel
character within the data. When a card fails
Fig 4.40
to be read correctly and needs to be swiped BCD character set used on track 2 and 3
again it is generally due to parity check or of most magnetic stripes.
LRC errors.
GROUP TASK Activity
Using the information above calculate the minimum width of the magnetic
stripe so that it is able to accommodate the maximum number of
characters on all three tracks. Compare your result with a real card.

GROUP TASK Activity

The data on track two of an ATM card contains a start sentinel followed
by the account number 12345678, a field separator, the encrypted PIN
2468 and finally an end sentinel and LRC character calculated over all
other characters. Produce the binary string stored on track 2 of this card.

All magnetic stripe readers contain a magnetic read head that operates using the same
principles as the read heads on tape drives and within hard disks. Some readers
require the user to swipe their card, whilst others require the card to be inserted into a
slot. Insertion style machines control the speed at which the magnetic stripe passes the
read head and hence tend to produce less errors. Such readers retain the card within
the machine until the transaction is completed. In ATMs insertion style readers are
used to increase security. For example failure to enter a correct PIN after a set number
of attempts or detecting that a card is stolen results in the card being retained within
the machine.
GROUP TASK Discussion
Brainstorm applications that use barcode readers and applications that use
magnetic stripe readers. Discuss likely reasons why each of these
applications uses one type of reader rather than the other.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 429

COLLECTION FROM FORMS

Forms are used to collect data required to process transactions. Forms can be paper
based where indirect users manually complete the form and then data is entered into
the system and subsequently batch processed at a later time. Common examples of
paper forms include Medicare forms, taxation returns, loan applications and
enrolment forms. Today many of these organisations also provide alternative web-
based data entry screens – this removes the need for participants to manually enter
data from paper forms into the system. Screens used for data collection are also forms.
These screens can be part of front-end client applications that connect via a local area
network to backend DBMS systems or they can be web-based clients where the data
travels over the Internet and then via a web server before it arrives for storage within
the system’s database. In either case the transactions can be processed immediately in
real time or they can be stored in a transaction file for later batch processing.
All forms are user interfaces; their purpose is to guide the user through the data
collection process such that the data is collected accurately and efficiently. Paper
forms are unable to react to user inputs whilst screens are able to provide real time
feedback in response to inputs. Data validation is used to improve accuracy. On
screens the validation criteria can be enforced. On paper forms validation criteria
cannot be enforced rather indicators of the required data are provided, for example
instructions, example data and input areas that restrict the length of data.
General form design principles
Some general principles that apply to paper, online and web-based form design
include:
• Know who the users are. What are their goals, skills, experience and needs? What
motivation has led the user to complete the form? Answers to these questions are
critical. The form must be usable given the ability of the users. The form will not
be completed if the user has little motivation, therefore the purpose of the forms
completion should be clear. Furthermore the purpose should reflect some user goal
or need. For example, a web-based form that requests personal details when the
user has no idea of how these details will be used or what they will receive in
return is unlikely to be completed honestly or at all.
• Identify the precise nature of all data items that will be collected. This includes the
data type, length and any other restrictions. Does entry of one data item determine
the possible values or alternatives for subsequent data items? Answers to such
questions help determine validation rules and the sequence of input fields.
• Consistency with other forms and applications. Capitalise on users past experience
and skills by using and arranging form components in familiar ways. For instance
on screens radio or option buttons should be round, whilst check boxes should be
square. On paper forms, a series of boxes is often used to control the number of
characters to be entered and to promote legible handwriting.
• Form components should be readable. Readability is affected by the actual words
and fonts used as well as the logical placement and grouping of related fields.
Underlined and capitalised text should be avoided; bold text is preferred where
extra emphasis is needed. Sans serif fonts are preferred – serif fonts are generally
reserved for large blocks of text.
• Forms should include significant areas of white space to visually imply grouping or
simply to rest the eyes. Colour and graphics should be used sparingly and only
when they improve the readability of the form. In general pastel background
colours are preferred with dark text and white input fields. Cluttered forms always
appear more complex compared to forms where elements are generously spaced.

Information Processes and Technology – The HSC Course

430 Chapter 4

Layout of labels and input fields

The layout and alignment of labels and input fields should lead the user through the
desired input sequence. In Western countries it is preferred for both labels and input
fields to be left justified. It is simpler to scan down a page when there is a hard line
down the left hand edge. Therefore all labels and input fields should be left justified.
Some common layouts are shown in Fig 4.41. The single column layouts on the left
are easier for the eye to scan down and each label is equally close to its corresponding
input field, however significantly more vertical space is used. The two column
designs require less vertical space, however the differing length of labels causes
problems. If the labels are left justified then all smaller length labels are positioned
some distance from their corresponding input fields. If the labels are right justified
then we have an undesirable jagged left edge. Introducing horizontal lines into the
design assists the eye to better link labels with their input fields; however including
such lines between all fields reduces the ability to scan downwards. The second single
column example groups input fields – groups that make logical sense should be
chosen. When designing forms compromises must be made. For large systems various
designs should be tested with many users before settling on a final layout.

Label Label:

Longer label:
Longer label
Much longer Label:

Option group label: � Option value 1

Much longer Label
� Option value 2
� Option value 3
Option group label
Main Action
� Option value 1
� Option value 2
� Option value 3
Label:
Main Action
Longer label:

Much longer Label:

Label Option group label: � Option value 1

� Option value 2
� Option value 3
Longer label
Main Action
Much longer Label

Label:
Another Label
Longer label:

And Another Much longer Label:

Option group label: � Option value 1

Option group label � Option value 2
� Option value 1 � Option value 3
� Option value 2
Main Action

Fig 4.41
Possible label and input field layouts.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 431

GROUP TASK Discussion

Identify and analyse other design features present on the example layouts
in Fig 4.41.

GROUP TASK Practical Activity

Create a suitable layout for collecting customer names, addresses, phone
numbers and email addresses.

Principles particular to the design of paper forms

Additional design considerations specific to paper-based forms include:
• Paper forms are used to collect data that will subsequently be input into a computer
system; therefore the paper form and the data entry screen need to be structured to
assist the data entry process as well as the manual completion of the paper form.
Paper forms should not merely be a printout of the corresponding data entry
screen; rather both versions should use the strengths of their respective mediums
whilst maintaining consistency in terms of the order of data elements.
• Paper-based forms cannot react to a user’s responses; hence instructions must be
available and clearly stated. General instructions relevant to the whole form should
be placed before the questions commence, whereas instructions for particular items
should be present at the point on the form where they are needed. For example if a
certain answer means the person must jump to question 9 then this needs to be
stated clearly; on a data entry screen the questions that are not needed can be
dulled or simply not displayed at all.
• Colour, texture, fonts and the paper itself cannot be altered when using paper
forms. Paper forms therefore should be designed so that these elements will work
for all, or at least the majority, of users. The paper should be thick enough that type
cannot be seen through the page. Consider having large print versions available for
sight-impaired users.
• Appropriate space for answers. The space provided for answers on a paper form
cannot increase or decrease; most people use the space provided as an indicator of
the amount of information they need to supply. On data entry screens it is possible
for such space to grow as needed, on paper forms such space needs to be more
carefully considered.
GROUP TASK Practical Activity
Obtain a copy of your school’s enrolment form. Analyse the design and
layout of this form and recommend areas for improvement.

Principles particular to the design of online screens

Additional considerations specific to the design of online screens that form part of
software applications include:
• Clearly show what functions are available. Users like to explore the user interface;
this is how most people learn new applications, therefore functions should not be
hidden too deeply. If a particular function is not relevant then it is better for it to be
dulled than for it to be hidden, this allows users to absorb all possibilities. At the
same time the user interface should not be overly complex.
• Every action by a user should cause a reaction in the user interface. This is called
feedback; without feedback that something is occurring, or has occurred, users will
either feel insecure or will reinitiate the task in the belief that nothing has

Information Processes and Technology – The HSC Course

432 Chapter 4

happened. Feedback can be provided in subtle ways; such as the cursor moving to
the next field, a command button depressing or the mouse pointer changing. Tasks
that take some time to complete should provide more obvious feedback indicating
the likely time for the task to complete.
• User actions that perform potentially dangerous changes should provide a way out.
Many software applications include an ‘undo’ feature, whilst others provide
warning messages prior to such dangerous tasks commencing. In either case the
user is given a method to reverse their action.
• Operating systems have their own standards for user interface design. These
standards should be adhered to wherever possible so that users knowledge and
skills can be transferred from other familiar applications.

GROUP TASK Discussion

Forms designed for touch screens are significantly different to those used
for keyboard and mouse entry. Identify the essential differences.

Principles particular to web forms

Additional considerations specific to the design of web-based forms include:
• The speed of individual user’s Internet connections is unknown. Therefore web-
based forms should try to validate data within the downloaded page wherever
possible – the aim being to reduce the amount of data transferred. If delays are
possible then feedback should be provided or processing delayed for later batch
processing.
• It is often possible to design a sequence of forms such that transmission of data
required for validation occurs prior to the next form in the sequence being
displayed. Commonly web forms validate all input fields together after a submit
button is clicked. If validation or other errors occur then users should be informed
of what the error is, why it occurred and how it can be rectified. Often the original
form is displayed again – if this technique is used then all correctly entered data
should be filled in rather then expecting all data to be re-entered. Furthermore the
data on one form can be used to determine the available options available on
subsequent forms.
• In general the hardware and software used to access web-based forms is largely
unknown. As a consequence particular care needs to be taken to ensure the
software technologies used will operate correctly on many different combinations
of hardware and software. In particular web pages should be tested within all
popular web browsers using a variety of different screen resolutions.
• Users are able to set their own preferences within web pages. Labels and input
fields can appear differently on different users machines even when they are using
identical hardware and software. Therefore web pages need to be designed so that
they will automatically format correctly based on the settings within each user’s
browser.
• Security of personal and other details is critical when using web-based forms.
Financial transaction data should always be encrypted during transmission. If
users are to feel confident divulging their details then the security measures used
during transmission and subsequent storage should be clear. For example, https,
which includes the secure sockets layer protocol should be used. Most browsers
display a small padlock to indicate that all data transferred will be securely
encrypted.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 433

GROUP TASK Practical Activity

Browse the web and examine a variety of different web-based forms.
Analyse each form and propose improvements.

Consider the design of the following forms:

Fig 4.42
Australian Taxation Office Short Tax Return for Individuals, page 1.

Information Processes and Technology – The HSC Course

434 Chapter 4

Fig 4.43
Main data entry screen from The UAI Estimator Version 10.0 for Windows.

Fig 4.44
Library search web-based form within Microsoft Internet Explorer.

GROUP TASK Discussion

Evaluate the design and layout of each of the above data collection forms.
In each case, propose possible improvements.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 435

ANALYSING DATA OUTPUT FROM TRANSACTION

PROCESSING SYSTEMS
Transaction processing systems contain large quantities of data that can be analysed to
improve the organisation’s performance. Past trends can be examined, the current
state of the organisation’s finances can be analysed and information can be used as
evidence to assist decision makers. Such analysis can be performed on the operational
data or on a data warehouse. In this section we consider data warehouses,
management information systems and decision support systems used to analyse
existing transaction data. Finally we consider enterprise systems, which are large
systems that perform critical tasks for an organisation.
Data Warehouse
We first considered data warehouses in chapter 2. A data warehouse is a large
database that includes historical copies of data from each of an organisation’s
operational databases. Data warehouses grow as new transaction data is added over
time. The data warehouse is not in itself an analysis tool rather it is a data resource
that analysis tools access to analyse the historical activities of the organisation.
Data warehouses are large snapshot copies of transaction databases, that is, they are
static or read only in nature. This means analysis can take place without concern over
simultaneous access or updating of transaction records. Furthermore the data
warehouse can act as an archive for the organisation’s historical data.
Advantages of data warehouses include:
• Old transaction data can be purged from the operational system and archived
within the data warehouse. This improves the performance of the operational
system, as less data needs to be examined during transaction processing.
• Analysis processes performed on the data warehouse do not degrade performance
of the operational system. Data warehouses are generally maintained on their own
hardware and software; hence they have no effect on the performance of the
operational systems.
• A data warehouse includes historical transaction data, often over 10 or more years.
Systems change completely and are regularly upgraded, however data warehouses
are designed such that all data is stored using a similar format. This common
format greatly simplifies analysis processes.
• Data warehouses are snapshot copies of the real data. This data does not and
should not change. Therefore analysis processes can proceed more efficiently.
There is no need to be concerned with record locks, ACID properties and data
integrity issues.
• Data warehouses centralise data from within the entire organisation. Commonly
this includes customer, sales, employee, payroll, production, marketing and any
other data created within an organisation. Having all such data in a central
repository means analysis can take place across the entire organisation.
• As a data warehouse is completely separate to the operational data it can be
organised differently to the operational data. For instance indexes can be created
on particular fields to improve the performance of analysis processes without risk
of degrading the operational system’s performance.
GROUP TASK Discussion
Often links are provided within the operational system that allow users to
view data within the data warehouse. Propose specific scenarios where
such access to historical data maybe an advantage.

Information Processes and Technology – The HSC Course

436 Chapter 4

Management Information Systems

A Management Information System (MIS) transforms data within transaction
processing systems into information to assist in the management of business
operations. MIS functions include the generation of sales reports, profit and loss
statements, graphs of sales trends and a variety of other reports required for the day-
to-day operation of organisations. Such reports are essentially summaries or statistical
analyses of existing data within the system. These reports are used by managers to
plan and direct the operation of the organisation.
In a small business such information is generated directly by the manager, whilst in
larger organisations one or multiple departments are dedicated to MIS processes. In
small systems the functions of the MIS are often contained within the transaction
processing system whilst in larger systems the MIS is a separate system or systems.
Large management information systems link to transaction data and perhaps to a data
warehouse. For instance reports that compare current productivity with historical
productivity require access to current transaction data and also to historical data
within the organisation’s data warehouse. Within large organisations MISs can
include one dedicated to generating information to assist financial management,
another to provide information to assist warehouse managers, another to assist
production managers and yet another to provide information to assist marketing
managers.
The participants who work within management information systems require strong
technical computer skills together with a solid grasp of business processes. These
personnel must transform the data within the system into information of relevance to
decision makers within their organisations. This can only occur when a mix of
technical IT skills and business knowledge is present.

Consider the following:

Each of the following is an example of information generated by an MIS. In each case

the data source is ultimately transaction data.
• A list of each product a factory produces together with the profit or loss made on
each over a 12 month period.
• A table listing each salesperson together with the total monthly value of their sales
over the past 12 months.
• The total value of cheques for each bank that pass through a large cheque clearance
facility on a particular day.
• A column graph displaying the total number of sick days taken by all employees on
each day of the week.
• A line graph for each product showing average total number sold each month over
a five year period.

GROUP TASK Discussion

For each of the above examples, identify the transaction data that has been
analysed to create the information.

GROUP TASK Discussion

For each of the above examples, discuss how the information could be
used by management to assist the operation of the organisation.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 437

Decision Support Systems

A Decision Support System (DSS), like an MIS, provides information to managers to
assist the decision making process. However decision support systems do much more
than merely summarise current transaction data. The analysis performed by decision
support systems presents possible solutions and is able to assess the likely
consequences of making particular decisions. For example an MIS creates a graph
summarising total sales made by each branch over the last month. A decision support
system is used to determine possible reasons why particular branches had higher or
lower sales totals. For instance, Online Analytical Processing (OLAP) systems (refer
page 224) are a type of online decision support system that allow decision makers to
drill down through different levels or dimensions within the data to uncover new
relationships and other information. The results can then be used to improve future
performance. In essence a decision support system can be thought of as an intelligent
kind of MIS. Many decision support systems look to the future, they are able to
generate forecasts and predictions based on historical or incomplete data. For example
predicting future interest rates or forecasting the weather are problems that do not
have a definite single correct solution. Decision support systems analyse the available
data to produce or suggest the most likely outcomes.
The second option in this course (Chapter 5) deals exclusively with decision support
systems, in this section we are concerned with how decision support systems are used
to analyse data generated by transaction processing systems. Be aware that not all
decision support systems analyse transaction data – there are various other possible
data sources.
Decision support systems that analyse transaction data commonly use a data
warehouse as their data source. Clearly the system’s hardware and software must be
capable of processing enormous amounts of data. Data mining is one decision support
technique that examines the raw data in an attempt to discover hidden patterns and
relationships. Data mining presents new information that was not originally intended
to be present within the data. Creating and querying data marts is another decision
support technique – data marts simplify and improve the efficiency of information
extraction from large data warehouses. A data mart is essentially a reorganised
summary of specific data from the data warehouse and/or transaction database. Each
data mart aims to meet specific decision support needs of a particular department. A
series of queries are executed either directly by users or via decision support software
to retrieve information that assists decision makers.

Consider the following:

To create a data mart select queries are run that create summaries of the data in the
transaction database or data warehouse and then the results of the query are used to
create a new table within the data mart. For example a query that returns the number
of each product sold per day could be used to create a new table. Within large data
warehouses that contain many millions or even billions of records the creation of the
new table will take some time – perhaps hours or even days. However this new data
mart table will be reused and as it contains far less data it can be analysed more
rapidly. Unfortunately whenever data is summarised some of the original detail (or
granularity) is lost. Therefore such summaries must be chosen carefully so that
required detail is retained.
Creating new tables for a data mart requires a corresponding reorganisation of the
database schema. This reorganisation aims to optimise the schema for decision
Information Processes and Technology – The HSC Course
438 Chapter 4

support processing – the original schema was designed to optimise transaction

processing. Often a simpler de-normalised schema based around one single large table
is preferred for decision support. Many of the reasons for normalising databases are
not present in data mart based decision support systems – existing data is never
altered and new data is added in bulk. Even when the large table contains a summary
of the raw data it can still include many millions of
records, therefore the schema should be designed so 1 m
that querying the large table only occurs when needed. m 1
One common strategy is to design the attributes of this
large table as a series of foreign keys to smaller tables. m
1
For instance a BranchID would be linked to a Branch m
table that details the location, region and state for each 1
branch. It is the detailed attributes within the smaller m
tables that are used within queries as the search, sort 1
and grouping criteria. Notice that such a schema forms
a star with the large table at the centre linking out to Fig 4.45
each of the smaller tables (refer Fig 4.45). Users are Typical star schema used for
able to efficiently identify criteria for queries by many data marts.
examining the smaller tables. The query is then
constructed using these criteria with joins to the larger table added later. Such a
simple schema allows users to quickly produce ad-hoc queries without the need to
understand the complexities of SQL statements needed to design queries with multiple
joins.
GROUP TASK Discussion
With regard to decision support systems, identify and list advantages of
data marts compared to data warehouses.

GROUP TASK Research

Research, using the Internet or otherwise, specific examples of business
decisions that have been made based on the analysis of historical
transaction data.

Consider the following:

A supermarket chain has some 200 stores across Australia. Each store’s transaction
database includes a record for each individual product scanned through a register for
each customer purchase. The chain’s head office creates a data mart for use by its
marketing department. Within this data mart a central table is created that contains a
single record for the total number of each product sold each day within each of the
200 stores.
GROUP TASK Discussion
Propose examples of information that can be retrieved from the above
data mart.

GROUP TASK Discussion

Identify and describe examples of information that CAN be derived from
the transaction database but CANNOT be derived from the data mart.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 439

Enterprise Systems
An enterprise is simply a large organisation, for example government departments,
large corporations and universities. An enterprise system is any system that performs
processes central to the overall operation of an enterprise. This includes critical
hardware, critical software applications and in particular critical data. For instance, a
typical university would have a variety of enterprise systems in operation, including a
student records system, a finance system, a payroll system, a human resources system
and also a content management system. Each of these enterprise systems is central to
the running of the university and operates throughout the university.

Consider the following enterprise system case study:

Dimension Data
Customer Size: 8600 employees
Organization Profile
Founded in 1983 and headquartered in Johannesburg, South Africa, Dimension Data is a global IT
provider and Microsoft Gold Certified Partner operating in 36 countries across five continents.
Business Situation
Dimension Data needed an enterprise-grade database that supported database mirroring for disaster recovery and
database snapshots for reporting.
Solution
Dimension Data upgraded its existing SAP R/3 infrastructure to Microsoft SQL Server 2005 Enterprise Edition running
on Microsoft Windows Server® 2003 Enterprise Edition operating system. The company moved to SQL Server 2005 to
take advantage of new features and enhanced functions of the database—including the Database Snapshot and
Database Mirroring features.
Dimension Data uses SQL Server 2005 Database Mirroring to maintain a continually updated copy of its data on a
separate server at each data center. It plans to expand its use of Database Mirroring to include storing a continuously
updated database at a geographically separate disaster recovery center.
The Database Snapshot feature of SQL Server 2005 is used for creating copies of the database throughout the day,
both for location backup and as a reporting database so that queries can be run without impacting the production
database.
A member of the HP Service Provider Program, Data Dimension supports its SAP infrastructure with HP ProLiant
servers equipped with Intel Xeon processors. Intel Xeon processors offer an ideal choice for demanding enterprise
applications such as SAP.
The SAP deployment architecture, which is identical for Johannesburg and London, includes:
o SAP R/3 data, totaling about 100GB s stored in a data warehouse running on SQL Server 2005 Enterprise Edition.
o Every three hours the Database Snapshot feature of SQL Server 2005 is used to create an updated copy of the SAP
database.
o SQL Server 2005 Analysis Services is used to create two multidimensional data cubes, to support faster data access
for analytics. The cubes are used by some analysts and other users.
o Dimension Data’s worldwide workforce accesses SAP information by logging into a portal supported by Microsoft
SharePoint® Portal Server. Microsoft Active Directory® directory service is used to help ensure information is
accessible on a role-based basis. SAP data is accessed by about 1,600 users.

Fig 4.46
Modified extract of Dimension Data case study (Source: microsoft.com)

GROUP TASK Discussion

Explain how the mirroring and snapshot features of Dimension Data’s
new enterprise solution protects their critical data..

GROUP TASK Discussion

SAP, HP and Microsoft are major players in the enterprise system market.
Research examples of enterprise systems that use these company’s product
i
Information Processes and Technology – The HSC Course
440 Chapter 4

SET 4E
1. Ferromagnetic materials used within MICR 6. In general, labels and input fields on forms
ink and toner: should be:
(A) is magnetically charged. (A) centred.
(B) can be magnetised. (B) right justified.
(C) are encoded with binary data. (C) left justified.
(D) are used during optical scanning. (D) fully justified.
2. Which of the following is true in regard to 7 Check digits and characters encoded on
the operation of barcode readers? magnetic stripes use:
(A) Light is reflected off the barcode to one (A) odd parity.
or more sensors. (B) even parity.
(B) Less light is reflected off dark colours. (C) checksums.
(C) The sensor(s) detect the intensity of (D) CRCs.
reflected light. 8 In regard to the design of paper forms, which
(D) All of the above. of the following is true?
3. In regard to the magnetic stripe on most (A) The input field order is determined by
ATM and credit cards, which of the the corresponding electronic data entry
following is true? form.
(A) The stripe contains 2 tracks, however (B) The form should make extensive use of
for most applications just one track colour and graphics to motivated users.
contains data. (C) All instructions should be included as a
(B) The stripe contains 3 tracks, however separate document.
for most applications just one track (D) Space for answers provides an indicator
contains data. of the amount of information required.
(C) The stripe contains 3 tracks, however 9 Designing forms such that they present well
for most applications just one track is in different fonts and screen resolutions is
read. particularly important when designing:
(D) The stripe contains 3 tracks, however (A) web forms.
for most applications two tracks are (B) paper forms.
read. (C) online forms.
4 Discovering hidden patterns and (D) forms within software applications.
relationships within large stores of data is 10. Which of the following reports is most likely
known as: to be produced by a DSS rather than a MIS?
(A) data mining. (A) Total sales by branch over the last 6
(B) data warehousing. months.
(C) decision support. (B) Average time to produce each product
(D) forecasting. during the last week.
5. MICR, barcode and magnetic stripe readers (C) Table detailing predicted profits
use which type of sensors respectively? resulting from different upgrade
(A) Magnetic, optical, magnetic. options.
(B) Optical, optical, magnetic. (D) Line graph displaying the total sales of
(C) Magnetic, magnetic, magnetic. a product for each month in the
(D) Optical, optical, optical. previous year.
11. Define the following terms:
(a) RFID (c) Magnetic stripe (e) Data mining (g) MIS
(b) Barcode (d) Data warehouse (f) DSS (h) Enterprise system
12. Describe the operation of each of the following collection devices?
(a) RFID reader (b) Barcode reader (c) Magnetic stripe reader
13. Contrast the design of paper-forms with the design of online/web forms.
14. A retailer sells personalised T-Shirts over the web. Customers upload their own image files,
which are subsequently printed on the T-Shirts. T-Shirts are available in four sizes - S, M, L and
XL. Cost is $30 for the first T-Shirt that uses a particular image and $20 for extra T-Shirts using
the same image. $15 is charged per order to cover postage and handling.
(a) Identify the data that needs to be collected to process a sale.
(b) Design a suitable data entry screen.
15. Distinguish between Management Information Systems and Decision Support Systems. Include
examples to illustrate your response.

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 441

ISSUES RELATED TO TRANSACTION PROCESSING SYSTEMS

There are numerous significant issues that should be considered when designing and
operating transaction processing systems. In this section we restrict our discussion to
issues in regard to:
• The changing nature of work.
• The need for alternative non-computer procedures.
• Bias in data collection.
• Data security, data integrity and data quality issues.
• Control and its implications for participants.

THE CHANGING NATURE OF WORK

The nature of work has seen significant change since the 1960s. These changes have
been both in terms of the types of jobs available and also in the way work is
undertaken. The widespread implementation of computer-based systems, and in
particular transaction processing systems, has been the driving force behind most of
these changes. In the early 1970s many thought that the consequence of new
technologies would be a reduction in the total amount of work needing to be done,
this has not occurred, rather new industries and new types of employment have been
created. Many people are now working longer hours, in more highly skilled and
stressful jobs than ever before.
Industries that once employed significant numbers of clerks have seen the greatest
changes. The majority of tasks traditionally performed by clerks are now automated.
Consider banks, transaction processing systems have largely replaced the numerous
clerks that once worked within each branch. Furthermore the widespread use of
ATMs, EFTPOS and credit cards mean customers rarely need to visit the bank. The
data entry tasks performed by bank staff are now performed by the customer, in the
case of ATMs and by retailers in the case of EFTPOS and credit card transactions. In
recent years the Internet has changed how transaction data is collected and processed.
It is now common to complete totally automated online purchases. No human
employed by the retailer needs to have any direct interaction with customers during
the transaction’s processing.
THE NEED FOR ALTERNATIVE NON-COMPUTER PROCEDURES
What happens when a transaction processing system fails? Perhaps there is a power
failure, lightning strike, fire, theft or communication lines are broken. Maybe the data
within the system has been lost or some hardware components are inoperable.
Recovery then involves purchasing replacement hardware, rebuilding systems and
restoring data. This takes time and during this time an alternative mode of operation is
required. For large centralised systems such problems are resolved by maintaining
backup power generators and redundant communication lines at complete mirrored
sites. For smaller systems, alternative non-computer procedures are needed if the
organisation is to continue to operate. Commonly the only alternative is a return to
paper based non-computer procedures whilst the problems are corrected.
Alternative non-computer procedures should be trailed and tested at regular intervals
to ensure they operate as planned. In particular such tests should ensure all
participants understand and are able to correctly implement the procedures. For
example when banks supply retailers with EFTPOS terminals they commonly supply
stock of manual paper forms. These paper forms allow the retailer to continue trading
despite failure of the EFTPOS system. However sales assistants must know how to
process sales using these paper forms – this requires training and regular testing.

Information Processes and Technology – The HSC Course

442 Chapter 4

Consider the following examples of system failure:

• A local post office is broken into and all computers are stolen. Upon phoning
Australia post it is determined that it will be one week before replacements arrive.
• A thunderstorm disrupts the communication lines into a large warehouse. The
warehouse is informed that the lines are unlikely to be restored for 3 days. The
transaction processing systems at the warehouse receives and processes hundreds
of orders per day that are subsequently shipped out by a fleet of 20 trucks.
• The ATMs outside a busy bank branch are ram raided and the cash boxes are
stolen. It will take at least two weeks for replacement ATMs to be installed.
GROUP TASK Discussion
Propose possible non-computer procedures that could be used to
minimise the effects of each of the above system failures.

GROUP TASK Discussion

Explain possible techniques that could be used to train participants and
test the procedures proposed above.

BIAS IN DATA COLLECTION

Bias is an inclination or preference that
influences most aspects of the collection Bias
process; the result of bias during An inclination or preference
collection is inaccurate data leading to towards an outcome. Bias
inaccurate outputs from the system. Those unfairly influences the
involved in collecting data must aim to outcome.
minimize the amount of bias present.
When deciding on the data to collect bias can be introduced. Often incomplete data is
collected with the aim of simplifying the system. For example it is common for loan
applications to collect data on a person’s income based entirely on their last few tax
returns. This data is used to assess each individual’s ability to repay the loan; the
assumption being that an individual’s income is likely to remain relatively constant
over time. In fact many people, particularly those who own or operate businesses, are
able to adjust their income to suit their expenses. By simply collecting past income
data the success of each loan application is biased in favour of salary and wage
earners at the expense of business owners.
Locating or identifying a suitable source of data for collection is another potential area
where bias can occur. Often efficiency of data collection means that the cheapest or
most available source of data is used rather than the best source of data. Consider
surveys; the data source for all surveys should aim to be a representative sample of the
entire population. However for ease of collection many organisations collect survey
data from users over the Internet. Internet users, in most cases, are not a representative
sample of the population; in general Internet users are younger, have higher incomes
and possess higher technology skills than the general population. Consequently results
derived from such surveys will not accurately reflect the entire population.
The collecting process itself should take into account the likely perceptions held by
those on whom the data is collected. People answer questions and fill out forms
differently based on their perception of how the data will be used. For example a
survey conducted by the Australian Taxation Office is likely to yield different results

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 443

to a similar survey conducted by the Australian Bureau of statistics. Individuals would

likely perceive the tax office as being interested in their individual responses whereas
a survey conducted by the Australian Bureau of Statistics is more likely to be viewed
as truly anonymous.
DATA SECURITY ISSUES
A summary of strategies we have examined to combat data security issues include:
• Passwords- Passwords are used to confirm that a user is who they say they are.
Once verified the user name is then used by the system to assign access rights.
• Backup copies- A copy of files is made on a regular basis.
• Physical barriers- Machines storing important data and information, or performing
critical tasks are physically locked away.
• Anti-virus software- All files are scanned to look for possible viruses. The anti-
virus software then either removes the virus or quarantines the file.
• Firewalls- A firewall provides protection from outside penetration by hackers. It
monitors the transfer of information to and from the network. Most firewalls are
used to provide a barrier between a local area network and the Internet.
• Data encryption- Data is encrypted in such a way that it is unreadable by those who
do not possess the decryption code.
• Audit trails- The transaction log includes details of who and when transactions
were performed. It is possible to work backwards and trace the origin of any
problem that may occur.
DATA INTEGRITY ISSUES
A summary of strategies we have examined to maximise data integrity include:
• Data validation- checks, at the time of data collection, to ensure the data is
reasonable and meets certain criteria.
• Data verification- regular checks to ensure the data collected and stored matches
and continues to match the source of the data.
• Referential integrity- ensuring all foreign keys in linked tables match a primary key
in the related table.
• ACID properties- ensuring transactions are never incomplete (atomicity), the data
is never inconsistent (consistency), transactions do not affect each other (isolation)
and that the results of a completed transaction are permanent (durability).
• Minimising data redundancy- Normalising reduces or eliminates duplicate data
within individual relational databases, however when transactions span multiple
databases issues will arise. The use of unique identifiers shared between
organisations allows individual entities to be accurately identified.
DATA QUALITY ISSUES
Data integrity is about the accuracy of the data – how well it matches and continues to
match its source. Data quality takes this one step further, it concerns how reliable and
effective the data is to the organisation. For example, responses on survey forms may
well be entered accurately into a system, however the quality of the data will be poor
if the respondents didn’t answer honestly or as intended. The resulting information
will be unreliable and ineffective. Other data quality issues occur when combining
data from different systems.
Consider creating a data warehouse from many databases. Some records will describe
the same entity differently; both may be correct, so which record is best? The
organisation of databases is likely to be different; different keys, data types or
schemas, for example. The meaning attached to an attribute can change over time;
perhaps a client application was modified and now stores different data in some old
Information Processes and Technology – The HSC Course
444 Chapter 4

field. Combining such data is difficult, unreliable and inefficient. Furthermore the
effectiveness and reliability of the information from subsequent data mining and
OLAP systems is reduced. Data Quality Assurance (DQA) standardises the definition
of data and includes processes that “scrub” or “cleanse” existing data so it meets these
data quality standards.

Consider the following data security, integrity and/or quality issues:

• A hacker gains access to an organisation’s system. They download customer credit

card details and use them to make various purchases over the Internet.
• An RTA employee alters driving test results so that licences are issued to people
who failed their driving test.
• A analyst using a data mining application uncovers links between sets of attributes
that cannot possibly be true.
• A bank customer determines that a funds transfer has not been completed. The
funds have left their account but have not been deposited into the other account.
GROUP TASK Discussion
For each of the above issues, determine the source of the issue and
suggest suitable strategies that would assist in preventing the issue
re-occurring in the future.

CONTROL AND ITS IMPLICATIONS FOR PARTICIPANTS

Control is the act of influencing or directing activities. In terms of managing the
activities of employees some level of control is reasonable. Management assigns tasks
and then quite legitimately expects employees to complete these tasks in a timely and
accurate fashion. However whenever one has control over another the relationship is
open to abuse. Determining precisely when control over participants is excessive is
often a grey area. Most would agree it is reasonable for managers to monitor the
activities of those they manage, however what level of monitoring is reasonable?
Should management control Internet access or be able to read all email messages? Is it
reasonable to monitor phone calls or remotely view a users desktop? Audit trails allow
management to track which records individuals have accessed; when is such tracking
reasonable? Answers to such questions differ considerably according to the
management style used and the nature of the tasks participants perform.
Current management theory suggests higher levels of productivity are achieved when
participants are motivated. Motivation improves when participants are given
responsibility for tasks and how they are completed. Motivated employees are less
likely to engage in undesirable activities and are much more likely to focus on work.
When employees are assigned boring or repetitive tasks they lose motivation and then
quite naturally seek to engage in other non-work related activities. When this occurs
management too often imposes authoritative controls such as excessive monitoring in
combination with negative consequences in an attempt to enforce control. Such
measures further reduce motivation resulting in even stricter controls being imposed –
a downward trend emerges. A more sustainable management style encourages trust
and motivates employees to take responsibility for work they complete.
GROUP TASK Discussion
As an employee, what level of monitoring by management do you feel
comfortable with? Brainstorm scenarios where monitoring and control of
participants is necessary (or at least justified)

Information Processes and Technology – The HSC Course

 Option 1: Transaction Processing Systems 445

CHAPTER 4 REVIEW
1. One operation within a transaction fails, 6. Which ACID property ensures either all or
what should occur? no operations within a transaction are
(A) Other operations within the transaction committed?
should be committed. (A) Atomicity
(B) The system should halt so that the (B) Consistency
reason for the failure can be corrected. (C) Isolation
(C) All operations within the transaction (D) Durability
should be rolled back. 7. Strict sequential processing of transactions
(D) No further transactions should be ensures which ACID property is observed?
performed until the problem is (A) Atomicity
resolved. (B) Consistency
2. Participants are those people who: (C) Isolation
(A) are the source of data used by the (D) Durability
system. 8. What is the main task performed by TPMs?
(B) receive information output from the (A) Providing an interface between many
system. transaction processing systems.
(C) interact directly with the system. (B) Ensuring transactions performed on a
(D) analyse data within the system. database observe the ACID properties.
3. Transaction logs used by most DBMSs (C) Monitoring and ensuring the security of
include details of records: transactions.
(A) prior to being altered. (D) Managing transactions that span
(B) after they have been altered. multiple database, systems and client
(C) added and deleted. applications.
(D) All of the above. 9. At most two sets of backups will be required
4. The file used to store data collected prior to to completely restore data when which of the
batch processing is commonly called: following backup types are used?
(A) an error file. (A) Full and incremental.
(B) a master file. (B) Full and differential.
(C) a database. (C) Incremental and differential.
(D) a transaction file. (D) Full backups only.
5. Checks to ensure data entered is reasonable 10. High speed MICR readers use which
are known as: technique to read the MICR line on cheques?
(A) data validation checks. (A) waveform
(B) data verification checks. (B) matrix
(C) data integrity checks. (C) CCD
(D) data redundancy checks. (D) LED

11. Provide at least TWO examples of systems where each of the following devices is used:
(a) MICR
(b) Barcodes
(c) Magnetic stripes
(d) RFID readers and tags
(d) TPMs
(e) Tape libraries
(f) Touch screens

Information Processes and Technology – The HSC Course

446 Chapter 4

12. Compare and contrast each of the following:

(a) User interfaces for real time processing with user interfaces for batch processing.
(b) Random (or direct) access with sequential access.
(c) Data validation with data verification.
(d) OLAP (Online Analytical Processing) and OLTP (Online Transaction Processing).
(e) Data warehouses and data marts.
(f) Data integrity with data quality.

13. (a) Recount the sequence of processes occurring to complete a typical credit card transaction.
Assume the transaction is initiated using an EFTPOS terminal supplied by the retailer’s bank.
(b) Describe different uses of transaction logs within transaction processing systems.
(c) Distinguish between the storage of collected data and the storage of processed data in a batch
transaction processing system using an example.

14. A company’s mail server records each email sent or received in a separate file. Incremental
backups using a round robin rotation occur automatically every hour to an online tape library. All
employees have full access to files within the tape library. Full backups are not made, however all
archive bits were set to true when the system was first installed. Tapes are changed every year as
there is sufficient capacity to store messages for 12 months.
(a) Critically evaluate the above backup procedure.
(b) Predict issues that may occur as a consequence of the above backup procedure.
(c) Propose and justify an improved procedure for backup and recovery.

15. Analyse an online web-based purchasing system of your choice.

(a) Determine the data items collected,
(b) Identify the operations performed during a purchasing transaction, and
(c) Evaluate the design of the data collection web forms.
(d) Explain how the company could analyse the collected data to identify areas for improving its
operations.

Information Processes and Technology – The HSC Course