Cloud Computing Introduction

and Foundations Concepts

Evolution towards Cloud Computing

Reference:
• [cam-san] Chapter 2
Cloud Computing - Definition
• NIST, the US National Institute of Standards and Technology: “Cloud
Computing is a model for enabling ubiquitous, convenient, on-demand
network access to a shared pool of configurable computing resources
that can be rapidly provisioned and released with minimal
management effort of service provider interactions”

• My Mom: “The place where my phone saves my emails, contacts,

pictures, so I don’t lose them”
Evolution of technology
• Cloud computing is not an abrupt innovation, instead, it is the result of a
series of developments that have taken place over the past few decades
• The concept traces back to 1960, when McCarthy (the same who coined
the term ‘artificial intelligence’) wrote:
• ‘computation may someday be organized as public utility’, like water, electricity, gas,
etc...
• It took, however, more than 40 years for technology to develop and mature
for cloud computing: only the composition of some key technologies has
enabled cloud computing
• Several decades of research, especially in the domain of parallel and
distributed computing, have paved the way
The origin – Mainframes
• Commercial usage of computing started around 1970s with
mainframes
• Organizations acquired such hardware to automate basic data
processing, e.g. payroll management
• Mainframes were large supercomputers installed in a room. They
were so high priced that it was hard for a company to have more than
one
• Multiple users were sharing the mainframe at the same time: the
common approach was to implement a time-shared utilization
• Different users were accessing the mainframe from terminals
The origin – Mainframes
• Terminals were dumb terminals,
i.e. were only capable of
transferring input-output data
to/from the mainframe (the
server)
• This centralized processing
environment was a bottleneck
• Users had always to wait for a
long time in queue
PC Revolution
• At the end of the 70s less
expensive computers came into
the market
• Those Personal Computers (PCs)
provided some processing and
storage capabilities to run
applications
• The need to maintain expensive
mainframes systems soon
became less important
Network of PCs
• Big organizations installed multiple PCs in different offices
• Each one, however, was functioning independently
• Network communication technologies were introduced to allow
communication between PCs for data transfer
• LAN: Local Area Network (LAN) and Wide Area Network (WAN) were
introduced at that time: computer in the same office or placed at
distance could communicate
• Over time communication technologies improved, increasing the
bandwidth
Distributed applications
• Client/Server computing: we • Peer to peer computing: each
have a server (offering some and every computer can play
service/information) and a the role of a client as well as a
client server

Interaction was limited to

exchange of small data due to
network limitations
Parallel processing
• In early 1980s it was believed that computer performance could only
be improved by producing faster and more powerful processors
• Parallel processing changed this idea: multiple processors (installed
on the same PC or on different PCs) working together to solve a single
(big) task that could not be executed on a single one
• The task must be designed to be parallel, i.e. composed by a set of
subtasks that could be executed in parallel
Faster Network Communication
• During 1980s advancements in
network communication improved
significantly data transfer rates (up to
100 Mbps for LANs and up to 64 Kbps
in WANs)
• Such advancements enabled
distributed computing systems: a
collection of processors interconnected
via communication network to execute
a complex task, impossible to be
executed on a single PC
 Distributed computing systems

• Applications are programmed

to be executed in a distributed
manner as a composition of
different tasks
• Each system has its own local
memory and processing power
• Nodes communicate to
exchange data or results
Cluster computing
• The next natural step was the creation of clusters: groups of multiple
nodes (computers) all connected to the same LAN to perform the same (or
similar) tasks
• Computers are clustered together for a more efficient computing (less
communication overhead) and to improve reliability.
• Pools of homogeneous
computing systems were
deployed to have more
processing power
• Among the computers of a
cluster, one node was selected
as cluster head to control the
cluster and distribute the tasks
Grid Computing
• The cluster head represented a single point of failure
• To mitigate this risk in early 90s a new architecture was introduced,
grid computing: a pool of resources in which the control
functionalities were implemented in a decentralized manner (not in
the cluster head)
• Computing nodes could be
deployed in the same network
or could be deployed in
different areas (even under
different administrative
domains)
Not enough
• Grid computing provided many advantages: a scalable architecture
that could offer high performance computing for complex tasks
• It had many drawbacks:
• Real-time scaling was not possible: more computing power/storage -> more
nodes -> change in the architecture
• The system was not fault tolerant: the failure of a node results in the failure of
a task
• Heterogeneous hardware required code adaptation
• Starting from the late 90s a new set of technologies matured, paving
the way for cloud computing
Hardware Virtualization
• The heterogeneous natures of computing systems architectures
posed a challenge to application (software) portability
• The execution of software on different architecture without changes
required to decouple software systems from the underlying physical
resources of computers
• This would allow to deal with hardware diversity but also to achieve
real time scalability
• For this reason, virtualization techniques were defined: a layer of
software over the hardware system that could simulate the whole
physical system environment
We will see that
extensively, don’t worry!
Web Based Technologies
• As computing technologies were evolving rapidly, users and business
were becoming more dependent on computer systems
• In addition to enabling grid computing to empower the development
of large computing systems, faster networks allowed users from
different geographical locations to collaborate in almost real time
• The World Wide Web arose as the killer application to spread
information and enable collaboration
• Web 2.0 appeared after 2002 as an evolution of Web 1.0, in which
user generated content was central
Service Oriented Architecture
• As the number of IT systems and their size increased, a new model for
software development was required
• Service Oriented Architecture (SOA) is a development method in
which applications are developed by leveraging software components
(software services) interacting each other
• Each service is an independent entity that communicate with the
others via messages
• Systems implemented using SOA remain flexible to changes

We will see that,

don’t worry!
Utility Computing Model
• Computing era reached a point in which we had:
• Scalable computing infrastructure made of heterogeneous resources
• Distributed computing environments empowered by high bandwidth
networks
• Collaborative environment across the world via web services
• Flexible applications via SOA
• Utility Computing: delivering of computing resources as utility,
following a pay per use model and on demand
• Crucial enabler for this model was hardware virtualization
 Cloud Computing … eventually

• Utility computing become

cloud computing
• Each one was essential,
from networking
advancements to novel
software development
methodologies
Timeline
2010: Microsoft
1970: DARPA’s TCP/IP 1999: Grid Computing Azure

1984: IEEE 802.3 1997: IEEE 2008: Google

Ethernet & LAN 802.11 (Wi-Fi) AppEngine
1966: Flynn’s Taxonomy
SISD, SIMD, MISD, MIMD 1989: TCP/IP
2007: Manjrasoft Aneka
IETF RFC 1122
1969: ARPANET
1984: DEC’s 2005: Amazon
1951: UNIVAC I, VMScluster AWS (EC2, S3)
First Mainframe
1975: Xerox PARC
Invented Ethernet 2004: Web 2.0
Clouds 1990: Lee-Calliau
1960: Cray’s First WWW, HTTP, HTML
Grids Supercomputer

Clusters

Mainframes

1950 1960 1970 1980 1990 2000 2010

Cloud computing today
• The first commercial initiative based on cloud computing concepts
was from salesforce.com in 1999
• The first large scale cloud computing service was commercialized by
Amazon in 2006 that lunched its Elastic Cloud Computing (EC2) and
Simple Storage Service (S3)
• Soon after other big players entered in the market with similar
services, e.g. Microsoft in 2009
• The term Cloud computing appeared with its present meaning for the
first time in 2006, used by Google CEO Eric Schmidt in a conference
Cloud computing in numbers
• The total expenditure worldwide reached a figure of $210 billion at
the end of 2020
• Demand for cloud services grows by 18% in 2019
• 90% of companies use some type of cloud service
• 80% of enterprises use Amazon Web Services as their primary cloud
platform
• There were 3.6 billion cloud users in 2018
• The positive impact of cloud technology is almost instantaneous. 80%
of companies report operation improvements within the first few
months of adopting the tech