International Journal of Engineering and Technical Research (IJETR)

ISSN: 2321-0869, Volume-2, Issue-11, November 2014

Distributed Versus Cloud Computing and data

security issues and new trends- Fog Computing
Sachin R. Desale, Kadambari V. Vanmali, Brajendra Singh Rajput

Abstract There is some confusion and misconception about A term Distributed System (DS) is a collection of
two terms, in this paper we are going to discuss about what are independent computers that appears to its users as a single
the similarities and differences between ditributed computing coherent system. It consists of collection of autonomous
and cloud computing. And furthermore we will discuss some computers, linked by a computer networks and equipped with
security isssues and new trends about the same, hope so this distributed system softwares.These softwares enables
paper will be helpfull to clarify the doubts about these two
computer to coordianate their activities and to share the
concepts or terms. Distributed computing comprises of multiple
software components that belong to multiple computers.
systems hardware, software and data too .
Whereas, a term Cloud computing is used to define a new class The term Distributed Operating System (DOS) can be
of computing that is based on network technology. defined as an Operating System which manages a collection
Cloud computing have been proposed different ways of of independent computers and make them to appear to its
using computers and to access and store our personal and users as a single sytem also, called Single System Image.
business information.Because of these new computing and
communication technology there arise new data security
challenges.Eventhough existing techniques use security
mechanisms, data theft attacks prevention fails, it has been
proved that these existing techniques are not enough to handle
such attacks.We propose a completely different approach for
securing data in the cloud using preventive decoy technology.
We monitor data access in the cloud and detect abnormal data
access patterns. When unauthorized access is suspected and then
verified using challenge questions. We launch a disinformation
attack by returning large amount of decoy information to the
attacker. This protects against the misuse of the users real data.

Index Terms Distributed computing, Cloud computing, Fig.1. Distributed Computing

Fog Computing, User behavior profiling, Decoy technology. The Open Software Foundations Distributed Computing
Environment (OSFS DCE) is a rich software technology that
enables the development of distributed applications across
I. INTRODUCTION heterogeneous systems, taking the advantage of open,
distributed computing. DCE is a suite of integrated services
Distributed Computing:- Distributed Computing is a
available from vendor-neutral source that enables
field of computer science that studies distributed
organizations to develop, use and maintain distributed
systems(DS). A distributed system is a software system in
applications across heterogeneous networks.It is used in
which components located on networked computers
exploratory systems, World Wide Web, stock markets, and
communicate and coordinate their actions by passing
telecommunication services. Following figure shows typical
messages. The components interact with each other in order to
architecture of DCE system.
achieve a common goal. Distributed computing comprises of
multiple software components that belong to multiple
computers. It is the system that works or runs as a single
system. The computers forming the distributed architecture
may or may not be closely located. These systems are often
preferred for the scalability factor. It is quite easy to add new
components to the system without disturbing the existing
system. A computer program that runs in a distributed system
is called a distributed program, and distributed

programming is the process of writing such

programs.There are many alternatives for the message
passing mechanism, including RPC-like connectors and
message queues. An important goal and challenge of
distributed systems is location transparency.
Fig. 2. DCE Architechture
Manuscript received November 18, 2014.

195 www.erpublication.org
 Distributed Versus Cloud Computing and data security issues and new trends- Fog Computing

Cloud Computing:- Cloud computing is entering our lives much as used and needed type of utility computing and the
and changing the way people consume information always on!, anywhere and any place type of network-based
dramatically. Clouds transform IT infrastructures with an computing.The flexibility and elasticity allows these
emphasis on making them flexible, affordable, and capable of systems to scale up and down at will utilising the resources of
serving millions of users, satisfying their computing or all kinds including CPU, storage, server capacity, load
storage demands. The design of early cloud computing balancing, and databases.
systems has evolved from, and was dominated by, the
concepts of cluster and grid computing. Currently, as the
concepts of the cloud become advanced and mature, cloud
networking and communication processes begin playing a
central role. Cloud Networking has emerged as a promising
direction for cost-efficient and reliable service delivery across
data communication networks. The dynamic location of
service facilities and the virtualization of hardware and
software elements are stressing the communication network
and protocols, especially when datacenters are interconnected
through the Internet.
Cloud computing is used to define a new class of
computing that is based on network technology. Cloud
computing takes place over the internet.It comprises of a
collection of integrated and networked hardware, software
and internet infrastructure.These infrastructures are used to
provide various services to the users.One of the biggest
advantage of using cloud computing is that it hides the
complexity and details of underlying infrastructure, and thus Fig. 3. Cloud System
users can easily use the services through simple graphical
interface.These systems are virtually managed and often The Three Layers/Models of Cloud Computing:-
distributed. Basically it is a step on from Utility Computing, As the delivery of IT resources or capabilities as a service
and a collection/group of integrated and networked hardware, is an important characteristic of Cloud Computing, the three
software and Internet infrastructure (called a platform).Using architectural layers of Cloud Computing are illustrated in
the Internet for communication and transport provides figure- 3 below.
hardware, software and networking services to clients. 1.) Infrastructure as a Service (IaaS) : IaaS offerings are
These platforms hide the complexity and details of the computing resources such as processing or storage which can
underlying infrastructure from users and applications by be obtained as a service. Examples are Amazon Web Services
providing very simple graphical interface or Application with its Elastic Compute Cloud (EC2) for processing and
Programming Interface(API). In addition, the platform Simple Storage Service (S3) for storage and Joyent who
provides on demand services, that are always on, anywhere, provide a highly scalable on-demand infrastructure for
anytime and any place.It is like Pay for use and as needed, runningWeb sites and rich Web applications.
elastic way of using technology.The hardware and software
services are available to general public, enterprises,
corporations and businesses markets.In short we can say,
Cloud computing is an umbrella term used to refer to Internet
based development and services.
A number of characteristics define cloud data, applications
services and infrastructure:
1. Remotely hosted: Services or data are hosted on remote
infrastructure.
2. Ubiquitous: Services or data are available from
anywhere.
3. Commodified: The result is a utility computing model
similar to traditional that of traditional utilities, like gas
and electricity - you pay for what you would want!
In other words we can define cloud computing as , it is a
shared pool of configurable computing resources, having
on-demand access and provisioned by service-providers. And
Public Clouds, Private Clouds , Community Clouds , Hybrid
Clouds these are the various types of cloud system. Massive
scale, Homogenity, Virtualization, Low Cost, Resilent
computing, Geographic distribution, Service Orientation, Fig.4. The Three Layers of Cloud Computing
On-demand service, Rapid elasticity, Broad network access,
Resource pooling, Measured service, these are very important 2.) Platform as a Service (PaaS): Platforms are an abstraction
characteristics of cloud computing which makes it quite layer between the software applications (SaaS) and the
different from other technologies.Cloud Computing is pay as virtualized infrastructure (IaaS). PaaS offerings are targeted

196 www.erpublication.org
 International Journal of Engineering and Technical Research (IJETR)
ISSN: 2321-0869, Volume-2, Issue-11, November 2014
at software developers. Developers can write their returned by the Cloud and delivered in such a way as to appear
applications according to the specifications of a particular completely legitimate and normal.
platform without needing to worry about the underlying The true user, who is the owner of the information, would
hardware infrastructure (IaaS). Developers upload their readily identify when decoy information is being returned by
application code to a platform, which then typically manages the Cloud, and hence could alter the Clouds responses
the automatic upscaling when the usage of the application through a variety of means, such as challenge questions, to
grows. Examples are the Google App Engine, which allow inform the Cloud security system that it has inaccurately
applications to be run on Googles infrastructure, and detected an unauthorized access. In the case where the access
Salesforces Force.com platform. is correctly identified as an unauthorized access, the Cloud
security system would deliver unbounded amounts of bogus
3.) Software as a Service (SaaS): SaaS is software that is information to the adversary, thus securing the users true data
owned, delivered and managed remotely by one or more from unauthorized disclosure. The decoys, then, serve two
providers and that is offered in a pay-per-use manner .Saas the purposes:
most visible layer of Cloud Computing for end-users, because (1) Validating whether data access is authorized when
it is about the actual software applications that are accessed abnormal information access is detected, and
and used. From the perspective of the user, obtaining software (2) Confusing the attacker with bogus information.
as a service is mainly motivated by cost advantages due to the
utility-based payment model, i.e. no up-front infrastructure
investment. Well known examples for SaaS offerings are II. SECURITY ISSUSES IN CLOUD COMPUTING
Salesforce.com and Google Apps such as Google Mail and
Google Docs and Spreadsheets. Cloud computing promises to significantly change the
way we use computers and access and store our personal and
New Trends:- Many proposals have been made to secure business information.With these new computing and
remote data in the Cloud using encryption and standard access communications paradigms arise new data security
controls. It is fair to say all of the standard approaches have Challenges.Existing data protection mechanisms such as
been demonstrated to fail from time to time for a variety of encryption have failed in preventing data theft attacks,
reasons, including insider attacks, mis-configured services, especially those perpetrated by an insider to the cloud
faulty implementations, and buggy code. Building a provider. Data theft attacks are amplified if the attacker is a
trustworthy cloud computing environment is not enough, malicious insider.This is considered as one of the top threats
because accidents continue to happen, and when they do, and to cloud computing by the Cloud Security Alliance (CSA).
information gets lost, there is no way to get it back.
The basic idea is that we can limit the damage of stolen data The Twitter incident is one example of a data theft attack
if we decrease the value of that stolen information to the from the Cloud. Several Twitter corporate and personal
attacker. We can achieve this through a preventive documents were ex-filtrated to technological website
disinformation attack. We posit that secure Cloud services TechCrunch , and customers accounts, including the account
can be implemented given two additional security features: of U.S. President Barack Obama, were illegally accessed.The
damage was significant both for Twitter and for its customers.
1. User Behavior Profiling:-It is expected that access to a While this particular attack was launched by an
users information in the Cloud will exhibit a normal means of outsider,stealing a customers admin passwords is much
access. User profiling is a well known technique that can be easier if perpetrated by a malicious insider. Rocha and
applied here to model how, when, and how much a user Correia outline how easy passwords may be stolen by a
accesses their information in the Cloud. Such normal user malicious insider of the Cloud service provider.
behavior can be continuously checked to determine whether
abnormal access to a users information is occurring. This We proposed a completley different approach to securing
method of behavior-based security is commonly used in fraud the cloud using decoy information technology , that we have
detection applications. Such profiles would naturally come to call Fog Computing.Here the meaning of Fog is
include volumetric information, how many documents are nothing but confusing to the attackers.We use this technology
typically read and how often. These simple user specific to launch disinformation attacks against malicious insiders
features can serve to detect abnormal Cloud access based or attackers , preventing them from distinguishing the real
partially upon the scale and scope of data transferred. sensitive customer data from fake worthless data.In this
paper, we propose two way of using Fog computing to
prevent attacks such as Twitter attack by deploying decoy
2. Decoys: Decoy information, such as decoy documents, information within the Cloud by the Cloud service customer
honeyfiles, honeypots, and various other bogus information and within personal online social networking profiles by
can be generated on demand and serve as a means of detecting individual users.
unauthorized access to information and to poison the thiefs
ex-filtrated information. Serving decoys will confound and III. DISRIBUTED VERSUS CLOUD ENVIRONMENT
confuse an adversary into believing they have ex-filtrated
useful information, when they have not.This technology may Following Table I will illustrate a complete and clear
be integrated with user behavior profiling technology to picture about similarities and differences between distributed
secure a users information in the Cloud. Whenever abnormal and cloud computing environrnment.
access to a cloud service is noticed, decoy information may be

197 www.erpublication.org
 Distributed Versus Cloud Computing and data security issues and new trends- Fog Computing

Parameters Distributed Computing Cloud Computing service provider when it comes to protecting their data. The
lack of transparency into, let alone control over, the Cloud
Definition Distributed computing Cloud computing is
comprises of multiple used to define a new
providers authentication, authorization, and audit controls
software components that class of computing only exacerbates this threat.
belong to multiple that is based on Much research in Cloud computing security has focused
computers. The system network technology. on ways of preventing unauthorized and illegitimate access to
works or runs as a single Cloud computing
system. Cloud computing takes place over the
data by developing sophisticated access control and
can be referred to as a form internet. It comprises encryption mechanisms. However these mechanisms have not
that originated from of a collection of been able to prevent data compromise. Van Dijk and Juels
distributed computing and integrated and have shown that fully homomorphic encryption, often
virtualization. networked hardware,
software and internet
acclaimed as the solution to such threats, is not a sufficient
infrastructure. data protection mechanism when used alone.

Goals 1.)Resource Sharing 1.) Reduced

2.)Openness Investments and
V. PROPOSED SYSTEM FOG COMPUTING
3.)Transparency Proportional Costs.
4.)Scalability 2.) Increased The proposed mechanism facilitates security features to
Scalability. data and thereby allows for detection of invalid access and
3).Increased
Availability and
thereby its prevention to enable valid distribution of data.
Reliability
a) User Behavior Profiling: Legitimate users of a computer
Types 1.)Distributed Computing 1.)Public Clouds system are familiar with the files on that system and where
Systems 2.)Private Clouds
2.)Distributed 3.)Community
they are located. Any search for specific files is likely to be
Information Systems Clouds targeted and limited. A masquerader, however, who gets
3.)Distributed Pervasive 4.)Hybrid Clouds access to the victims system illegitimately, is unlikely to be
Systems familiar with the structure and contents of the file system.
Their search is likely to be widespread and untargeted.
Characteristics 1.) A task is distributed 1.) It provides a
amongst different machines shared pool of Based on this key assumption, we profiled user search
for the computation job at the configurable behavior and developed user models trained with a oneclass
same time. computing resources. modeling technique, namely one-class support vector
2.) Technologies such as 2.) An on-demand machines. The importance of using one-class modeling stems
Remote Procedure calls and network model is
Remote Method Invocation used to provide from the ability of building a classifier without having to share
are used to construct access. data from different users. The privacy of the user and their
distributed computations. 3.) The clouds are data is therefore preserved.
provisioned by the We monitor for abnormal search behaviors that exhibits
Service Providers.
4.) It provides broad
deviations from the user baseline.Accoording to our
network access. assumption, such deviations signal a potential masquarade
Cons 1.) Higher level of failure of 1.) More elasticity attack.Our experiments on local file setting validate our
nodes than a dedicated means less control assumption and demonstrated that we could reliably detect all
parallel machine. especially in the case
2.) Few of the algorithms are of public clouds.
simulated masquarade attacks using this approach with a very
not able to match with slow 2.) Restrictions on low false positive rate of 1.12%.
networks. available services
3.) Nature of the computing may be faced, as it b) Decoy Technology: We placed traps within the file
job may present too much depends upon the
system. The traps are decoy files downloaded from a Fog
overhead. cloud provider.
computing site, an automated service that offers several types
of decoy documents such as tax return forms, medical records,
Table I. Distributed versus Cloud Computing credit card statements, e-bay receipts, etc.The decoy files are
downloaded by the legitimate user and placed in
IV. EXISTING SYSTEM FOR CLOUD SECURITY highly-conspicuous locations that are not likely to cause any
interference with the normal user activities on the system. A
The existing mechanisms only facilitate security features masquerader, who is not familiar with the file system and its
to data and thereby dont allow for detection of invalid access contents, is likely to access these decoy files, if he or she is in
and thereby its prevention to enable valid distribution of data. search for sensitive information, such as the bait information
Businesses, especially startups, small and medium businesses embedded in these decoy files. Therefore, monitoring access
(SMBs), are increasingly opting for outsourcing data and to the decoy files should signal masquerade activity on the
computation to the Cloud. system.The advantages of placing decoys in a file system are
This obviously supports better operational efficiency, but threefold:
comes with greater risks, perhaps the most serious of which
are data theft attacks.Data theft attacks are amplified if the (1) the detection of masquerade activity
attacker is a malicious insider. This is considered as one of the (2) the confusion of the attacker and the additional costs
top threats to cloud computing by the Cloud Security incurred to distinguish real from bogus information.
Alliance. While most Cloud computing customers are
well-aware of this threat, they are left only with trusting the

198 www.erpublication.org
 International Journal of Engineering and Technical Research (IJETR)
ISSN: 2321-0869, Volume-2, Issue-11, November 2014
(3) the deterrence effect which, although hard to measure, operations or not. Accordingly, the system will decide
plays a significant role in preventing masquerade activity by whether the user should be redirected to the decoy
risk-averse attackers. environment.

We posit that the combination of these two security features 4. Data Access Module:
will provide unprecedented levels of security for the Cloud. The data available for user access will be authenticated using
No current Cloud security mechanism is available that a separate user key specified by the application to the user
provides this level of security.Experiments conducted in a during registration. Based on the validity of this user key the
local file setting provide evidence that this approach may system will redirect the user to the Decoy Module for tracking
provide unprecedented levels of user data security in a Cloud and prevent invalid distribution of data. This key data will be
environment. secured in the system using HMAC mechanism.

VI. ALGORITHMS 5. Decoy Module:

HMAC Algorithm for Key Security:- This standard This module will facilitate the system to redirect invalid users
describes a keyed-hash message authentication code to a dummy set of modules wherein invalid data will be
(HMAC), a mechanism for message authentication using distributed to the invalid user and the user activities will be
cryptographic hash functions. HMAC can be used with any notified to the admin. Thus, the system will not notify the
iterative Approved cryptographic hash function, in invalid user about the detection of invalid activity and prevent
combination with a shared secret key. This standard specifies further attack on the system.
an algorithm for applications requiring message
authentication. Message authentication is achieved via the
construction of a message authentication code (MAC). VIII. COMBINING THE TWO TECHNIQUES
MACs based on cryptographic hash functions are known as The correlation of search behavior anomaly detection with
HMACs. trap-based decoy files should provide stronger evidence of
The purpose of a MAC is to authenticate both the source of malfeasance, and therefore improve a detectors accuracy.
a message and its integrity without the use of any additional We hypothesize that detecting abnormal search operations
mechanisms. HMACs have two functionally distinct performed prior to an unsuspecting user opening a decoy file
parameters, a message input and a secret key known only to will corroborate the suspicion that the user is indeed
the message originator and intended receiver.Additional impersonating another victim user. This scenario covers the
applications of keyed-hash functions include their use in threat model of illegitimate access to Cloud data.
challenge-response identification protocols for computing Furthermore, an accidental opening of a decoy file by a
responses, which are a function of both a secret key and a legitimate user might be recognized as an accident if the
challenge message.An HMAC function is used by the search behavior is not deemed abnormal.In other words,
message sender to produce a value (the MAC) that is formed detecting abnormal search and decoy traps together may make
by condensing the secret key and the message input. a very effective masquerade detection system. Combining the
The MAC is typically sent to the message receiver along two techniques improves detection accuracy.
with the message. The receiver computes the MAC on the We use decoys as an oracle for validating the alerts issued
received message using the same key and HMAC function as by the sensor monitoring the users file search and access
was used by the sender, and compares the result computed behavior. In our experiments, we did not generate the decoys
with the received MAC. If the two values match, the message on demand at the time of detection when the alert was issued.
has been correctly received, and the receiver is assured that Instead, we made sure that the decoys were conspicuous
the sender is a member of the community of users that share enough for the attacker to access them if they were indeed
the key. trying to steal information by placing them in highly
conspicuous directories and by giving them enticing names.
VII. TECHNOLOGY MODULES With this approach, we were able to improve the accuracy of
1. User Authentication: our detector. Crafting the decoys on demand improves the
The user is facilitated here to authenticate and thus, ensure accuracy of the detector even further. Combining the two
that only valid users can access the application. But, it also techniques, and having the decoy documents act as an oracle
tracks the user login operation and accordingly redirects the for our detector when abnormal user behavior is detected may
user to the decoy application. lower the overall false positive rate of detector.

2. Admin Module: We trained twenty classifiers with computer usage data

This module facilitates the admin to manage users, the data from 10 computer science students collected over a period of
stored and the invalid activities occurring within the 5 days on average. The classifiers were trained using the
application. Thus, this user will be responsible for tracking search behavior anomaly detection technique .We also
the application functionalities. A set of valid access rules will trained another 10 classifiers using a detection approach
also be defined by the admin for identification of invalid that combines user behavior profiling with monitoring
users. access to decoy files placed in the local file system, as
described above. We tested these classifiers using simulated
3. File Search Module: masquerader data. Figure 5 displays the AUC scores achieved
This module will enable to track whether the search by both detection approaches by user model. The results show
operations executed by the user follow a valid set of that the models using the combined detection approach

199 www.erpublication.org
 Distributed Versus Cloud Computing and data security issues and new trends- Fog Computing

achieve equal or better results than the search profiling unauthorized data access or exposure is suspected, and later
approach alone. verified, with challenge questions for instance, we inundate
the malicious insider with bogus information in order to dilute
AUC Comparison by User the users real data.

ACKNOWLEDGMENT
This work was partially supported by the European Union
FP7/2007-2013 under project TCLOUDS (grant agreement
257243), and based on work supported by the Defense
Advanced Research Projects Agency (DARPA) under the
ADAMS (Anomaly Detection at Multiple Scales) Program
with grant award number W911NF-11-1-0140 and through
the Mission-Resilient Clouds (MRC) program under Contract
FA8650-11-C-7190.

REFERENCES
[1] Cloud Security Alliance, Top Threat to Cloud Computing
V1.0,March 2010. [Online]. Available :
https://cloudsecurityalliance.org/topthreats/csathreats.v1.0.pdf.
[2] M. Ben-Salem and S. J. Stolfo, Combining a baiting and a user
search profiling techniques for masquerade detection, in Columbia
UniversityComputer Science Department, Technical Report #
User Number cucs-018-11, 2011.
https://mice.cs.columbia.edu/getTechreport.php?techreportID=1468.
[3] The permanent and official location for the Cloud Security Alliance
Fig. 4. AUC Comparison by User model for the Search Top Threats research is:
Profiling and Combined Approach http://www.cloudsecurityalliance.org/topthreats
The Keyed-Hash Message Authentication Code (HMAC).pdf
The results of our experiments suggest that user profiles [4] International Journal of Computer Networks & Communications
(IJCNC) Vol.5, No.1, January 2013 DOI : 10.5121/ijcnc.2013.5112171
are accurate enough to detect unauthorized Cloud access . A USER PROFILE BASED ACCESS CONTROL MODEL
When such unauthorized access is detected, one can respond [5] FOG Computing Mario Nemirovsky ICREA/BSC With the
by presenting the user with a challenge question or with a Coloaboration Rodolfo Milito CISCO Marcelo Yanuzzi UPC.
decoy document to validate whether the access was indeed [6] A. Bessani, M. Correia, B. Quaresma, F. Andre, and P. Sousa.
DepSky: Dependable and secure storage in a cloud-of-clouds.
Unauthorized, similar to how we used decoys in a local file
In Proceedings of the European Conference on Computer
setting, to validate the alerts issued by the anomaly detector Systems (EuroSys), pages 3146, April 2011.
that monitors user file search and access behavior. [7] Cloud Security Alliance. Top threats to cloud computing v1.0,
March 2010.
[8] D. Nurmi, R. Wolski, C. Grzegorczyk, G. Obertelli, S.
Abbreviations and Acronyms
Soman, L. Youseff, and D. Zagorodnov. The Eucalyptus
Opensource cloud-computing system. In Proceedings of the IEEE
AUC-Authentication Certificates, API-Application International Symposium on Cluster Computing and the
Programming Interface, CSA-Cloud SecurityAlliance, Grid pages 124131, 2009.
DCE-Distributed Computing Environment, DS-Distributed [9] E. Grosse, J. Howie, J. Ransome, J. Reavis, and S. Schmidt.
Cloud computing roundtable. IEEE Security Privacy, 8(6):17
Systems, HMAC- Hash Message Authentication Code, 23, 2010.J. Jones. (1991, May 10). Networks (2 nd
SMB-Small and Medium Business etc. ed.)[Online].Available: http://www.atm.com

IX. CONCLUSION
In this paper, we discussed about similarities and
differences regarding to two concepts, distributed computing
and cloud computing. Distributed computing is a computing
concept that, in its most general sense, refers to multiple
computer systems working on a single problem. In distributed
computing, a single problem is divided into many parts, and
each part is solved by different computers.And Cloud
computing is an umbrella term used to refer to Internet based
development and services.
And in this paper, we also present a novel approach to
securing personal and business data in the Cloud. We propose
monitoring data access patterns by profiling user behavior to
determine if and when a malicious insider illegitimately
accesses someones documents in a Cloud service. Decoy
documents stored in the Cloud alongside the users real data
also serve as sensors to detect illegitimate access. Once

200 www.erpublication.org