+ All Categories
Home > Documents > [IEEE 2012 International Conference on Communication Systems and Network Technologies (CSNT) -...

[IEEE 2012 International Conference on Communication Systems and Network Technologies (CSNT) -...

Date post: 06-Jan-2017
Category:
Upload: bhavna
View: 215 times
Download: 3 times
Share this document with a friend
6
Trustworthy Service Provider Selection in Cloud Computing Environment Punam Bedi Department of Computer Science University of Delhi -110007, Delhi, India [email protected] Harmeet Kaur Department of Computer Science Hans Raj College University of Delhi -110007, Delhi, India [email protected] Bhavna Gupta Department of Computer Science University of Delhi Delhi, India [email protected] AbstractCloud computing is an emerging computing paradigm which allows sharing of massive, heterogeneous, elastic resources among users. Despite of all the hype surrounding the cloud, users are still reluctant to adopt cloud computing because public cloud services process users’ data on machines that users do not own hence there is a fear of leakage of users’ commercially sensitive data. Due to these reasons, it is very necessary that cloud users’ be vigilant while selecting the service providers present in the cloud. To address this problem, selection of trustworthy cloud providers is proposed where users or enterprises employ the services of trustworthy service providers in the cloud. The paper proposes a system based on cooperative model of society where users select trustworthy service providers based on the recommendations given by their trustworthy acquaintances. The uncertainty present in the recommendations is handled through Fuzzy Inference System (FIS). Fuzzy inference system is capable of inferring crisp output even when the inputs are imprecise or uncertain, mimicking the reasoning of human mind. Experiments confirm that selection of trustworthy cloud providers is an effective and feasible way of estimating the trustworthiness of the service providers and thus helping users in protecting their data. Keywords- Trust; Reputation; Agent; Recommendation; Fuzzy Inference Engine I. INTRODUCTION Cloud computing is a new paradigm where applications are delivered as services over the internet using the hardware and system software of the datacenters. Its goal is to regard “computing power” as the public utility like water, electricity which meets personal and social needs with low cost and high efficiency [16]. Cloud computing provides various forms of cloud services resources, such as web services, application programming interface (API), infrastructure as a service (IaaS), platform as a service (PaaS) and software as a service (SaaS) to users. It integrates and divides service resources to meet requirements of individual users bringing scale advantage, reduce costs and quick response to requirements. Cloud services are massive, heterogeneous, distributed and it may happen that because of unjustifiable interest of service providers, they may provide incomplete, false or even vicious service descriptions, which may mislead the users and make the cloud environment open, uncertain and deceptive for users. Even the SLA (Service- Level Agreement) signed by interacting parties specifying the legislation, internally or externally imposed policies that will be followed during that service has no connotation because there is no method to measure that those written policies have been practiced by service providers or not [16]. Hence it is very difficult in the cloud environment to find services that meet user’s requirements and trustworthy enough to harness the privacy of the user’s data. Apprehending this foremost problem for the cloud computing environment, it is certain that there is a requirement of scalable and dynamic method, for users to select service provider from plethora of choices available in the market. As it happens in our society, people like to do transactions with trustworthy entities to safeguard themselves from the damage [15], similarly if somehow users are able to compute trustworthiness of service providers before using his services then risk of misuse of deployed application and data of users can be secured. Combined with multi agent technology, this paper proposes a framework for the selection of trustworthy service provider present in cloud environment where primarily two types of agents exist in the system - user agents and service provider agents. User agents select the trustworthy service provider agents for its application using social recommendation process. For this the user agent forms the query from user’s requirements and takes the recommendations about different service providers from its trustworthy acquaintances by sending them the query. The requested trustworthy acquaintances may further take recommendations from their acquaintances and process will go on till T timeout permits. T timeout is the variable which stores the time for which the user can wait for the result and is specified by the user only. All the trustworthy acquaintances who have been asked for recommendations give the aggregated recommendations i.e. combination of their direct experience with the service providers and recommendations collected from their trustworthy acquaintances, as attributes of the service provider in the scale of 0-9, to the requesting agent. The recommendations collected by the user agent, using the 2012 International Conference on Communication Systems and Network Technologies 978-0-7695-4692-6/12 $26.00 © 2012 IEEE DOI 10.1109/CSNT.2012.158 712 2012 International Conference on Communication Systems and Network Technologies 978-0-7695-4692-6/12 $26.00 © 2012 IEEE DOI 10.1109/CSNT.2012.158 717 2012 International Conference on Communication Systems and Network Technologies 978-0-7695-4692-6/12 $26.00 © 2012 IEEE DOI 10.1109/CSNT.2012.158 714
Transcript

Trustworthy Service Provider Selection in Cloud Computing Environment

Punam Bedi Department of Computer Science

University of Delhi -110007, Delhi, India

[email protected]

Harmeet KaurDepartment of Computer Science

Hans Raj College University of Delhi -110007, Delhi,

India [email protected]

Bhavna GuptaDepartment of Computer Science

University of Delhi Delhi, India

[email protected]

Abstract— Cloud computing is an emerging computing paradigm which allows sharing of massive, heterogeneous, elastic resources among users. Despite of all the hype surrounding the cloud, users are still reluctant to adopt cloud computing because public cloud services process users’ data on machines that users do not own hence there is a fear of leakage of users’ commercially sensitive data. Due to these reasons, it is very necessary that cloud users’ be vigilant while selecting the service providers present in the cloud. To address this problem, selection of trustworthy cloud providers is proposed where users or enterprises employ the services of trustworthy service providers in the cloud. The paper proposes a system based on cooperative model of society where users select trustworthy service providers based on the recommendations given by their trustworthy acquaintances. The uncertainty present in the recommendations is handled through Fuzzy Inference System (FIS). Fuzzy inference system is capable of inferring crisp output even when the inputs are imprecise or uncertain, mimicking the reasoning of human mind. Experiments confirm that selection of trustworthy cloud providers is an effective and feasible way of estimating the trustworthiness of the service providers and thus helping users in protecting their data.

Keywords- Trust; Reputation; Agent; Recommendation; Fuzzy Inference Engine

I. INTRODUCTION Cloud computing is a new paradigm where applications are delivered as services over the internet using the hardware and system software of the datacenters. Its goal is to regard “computing power” as the public utility like water, electricity which meets personal and social needs with low cost and high efficiency [16]. Cloud computing provides various forms of cloud services resources, such as web services, application programming interface (API), infrastructure as a service (IaaS), platform as a service (PaaS) and software as a service (SaaS) to users. It integrates and divides service resources to meet requirements of individual users bringing scale advantage, reduce costs and quick response to requirements. Cloud services are massive, heterogeneous, distributed and it may happen that because of unjustifiable

interest of service providers, they may provide incomplete, false or even vicious service descriptions, which may mislead the users and make the cloud environment open, uncertain and deceptive for users. Even the SLA (Service-Level Agreement) signed by interacting parties specifying the legislation, internally or externally imposed policies that will be followed during that service has no connotation because there is no method to measure that those written policies have been practiced by service providers or not [16]. Hence it is very difficult in the cloud environment to find services that meet user’s requirements and trustworthy enough to harness the privacy of the user’s data. Apprehending this foremost problem for the cloud computing environment, it is certain that there is a requirement of scalable and dynamic method, for users to select service provider from plethora of choices available in the market.

As it happens in our society, people like to do transactions with trustworthy entities to safeguard themselves from the damage [15], similarly if somehow users are able to compute trustworthiness of service providers before using his services then risk of misuse of deployed application and data of users can be secured. Combined with multi agent technology, this paper proposes a framework for the selection of trustworthy service provider present in cloud environment where primarily two types of agents exist in the system - user agents and service provider agents. User agents select the trustworthy service provider agents for its application using social recommendation process. For this the user agent forms the query from user’s requirements and takes the recommendations about different service providers from its trustworthy acquaintances by sending them the query. The requested trustworthy acquaintances may further take recommendations from their acquaintances and process will go on till Ttimeout permits. Ttimeout is the variable which stores the time for which the user can wait for the result and is specified by the user only.

All the trustworthy acquaintances who have been asked for recommendations give the aggregated recommendations i.e. combination of their direct experience with the service providers and recommendations collected from their trustworthy acquaintances, as attributes of the service provider in the scale of 0-9, to the requesting agent. The recommendations collected by the user agent, using the

2012 International Conference on Communication Systems and Network Technologies

978-0-7695-4692-6/12 $26.00 © 2012 IEEE

DOI 10.1109/CSNT.2012.158

712

2012 International Conference on Communication Systems and Network Technologies

978-0-7695-4692-6/12 $26.00 © 2012 IEEE

DOI 10.1109/CSNT.2012.158

717

2012 International Conference on Communication Systems and Network Technologies

978-0-7695-4692-6/12 $26.00 © 2012 IEEE

DOI 10.1109/CSNT.2012.158

714

above process contains a lot of uncertainty and fuzziness. To handle it user agent applies fuzzy inference system (FIS) and obtains the reputation of the various service providers recommended by the trustworthy acquaintances. Fuzzy inference system is capable of inferring crisp output even when the inputs are imprecise or uncertain, mimicking the reasoning of a human mind. This paper constructs a quantitative trust model for cloud computing environment based on the interactions of the system users with the service providers which helps the users in service provision and discovery. Once the services of cloud service provider from the recommended list are used by the cloud user, trust is updated on recommenders , based on the importance of their recommendations.

Rest of the paper is organized as follows. Section 2 details related work. Trust based cloud computing for selection of trustworthy service providers framework is discussed in section 3 followed by experimentation results in section 4. Conclusion is presented in section 5.

II. Related work Trust and reputation based system have already been used to differentiate the service providers based on quality of services in various fields i.e. e-commerce, grid computing, P2P etc. [1][9][17]. With the growing number of challenges in the cloud computing, it is pointed out that there is a need of method for selecting service providers [16]. The method to accredit the cloud providers using third party assurance body is given in [5] to reduce the problems relating to data security and privacy. In [11], service quality parameters are used to evaluate the service providers but uncertainty present in the parameters evaluation is not handled which is taken in our proposed framework using FIS. Few numbers of articles also handles security issues, with recommendations in cloud computing [6][8]. Recently trust based reputation system is proposed in inter cloud environment [12] but quality parameters for evaluating the service providers are not considered. The proposed work in this paper can be taken as continuation to [14][15] with Fuzzy inference engine used to handle the uncertainty in recommended parameters. In the proposed work recommendations from trustworthy acquaintances are taken about the attributes of those service providers whose services have been used by them and fuzzy inference system is applied on those recommendations to get the reputation of various service providers present in the cloud system. .

III. SELECTION OF TRUSTWORTHY SERVICE PROVIDERS IN CLOUD COMPUTING ENVIORNMENT (THE FRAMEWORK)

The essence of the cloud computing is to provide cloud services by network where users delegate their applications to service providers to save time, cost and to get quality results. Because of plethora of choices available on internet about service providers, it is not easy to take decision [12]. The decision, of choosing a service provider among several, is far difficult than deciding about the products available in

e-marketplace. Since, products have many tangible cues attached like style, color, hardness, label, package etc. whereas most of these characteristics are absent in case of services provided by service providers. Due to intangible nature, services cannot be counted, measured, tested in advance of sales [4]. In the absence of tangible cues attached with services, while selecting service providers for a given service several other factors like trust, reputation needs to be addressed. The definition of the reputation taken in this paper is as follows “Reputation is a value corresponding to each of the entity present in the cloud environment based on the trust exhibited by it in the past” [2] i.e. if a service provider is known to provide certain good qualities of services over a period of time irrespective of its limitations, then it is assumed to have a good reputation. As in a cloud environment, data centers may be geographically distributed and owned by the different individuals so a multi agent environment is proposed for the framework. The multi agent system proposed is organized in such a way that user and cloud provider present in the cloud is accompanied by an agent. In this proposed framework, there are two types of agents: 1. User agent: User of the cloud submits his application to

its corresponding user agent with specified requirements e.g. workload, execution deadline, budget limit etc. It is now the responsibility of the user agent to select a suitable service provider for the user satisfying the requirements specified by her/him. Here in proposed system, the user agents act as a community and help each other at the time of request of any user’s query. Every agent maintains a database in which besides their personal experience about service providers, trust on various trustworthy agents is also stored.

2. Service provider agent: Every cloud owner communicates through his agent called service provider agent. The data corresponding to every resource present in the data center such as its availability, its workload is maintained by this agent.

Basic methodology of the system consists of the following steps: 1. Generation of a query from the requirements of the

user. 2. Generation of recommendations for the user. 3. Updating trust level on trustworthy acquaintances

A. Generation of a query from the requirements of the user Whenever a user has some application to be given to

cloud then he will submit his application with his set of requirements to the user agent such as type of services required, budget, response time (time for which the user can wait for the result of execution) etc. It is now the responsibility of the user agent to select service providers with required services for the job, meeting the specific requirements of the user and give the result of job execution within the specified time to user. To accomplish this, the user agent prepares a query from those set of requirements which is sent to trustworthy acquaintances to get recommendations.

713718715

B. Generating the recommendations for the user In our framework users agents exist in a network that acts

as a society and follow social recommendation process. As we know in our society to make decisions people generally use their own experience and take recommendation from their trustworthy acquaintances which act as referral recommendations. For example if A trusts B and B trusts C and if A requests something from his trustworthy acquaintance B, B can also take help from his trustworthy acquaintance C to give recommendations. So finally A has referral recommendations from those persons who are unknown to him directly i.e. C. Similarly in our system at the time of request, the user agent take recommendation from those known to it, they may further take recommendations from those known to them and so on. Let xij represents the trust of agent i on agent j and xji represents trust level of agent j on agent i. Also it is not necessary that xij = xji. These trustworthy relationships form a web of trust and the advantage of this web is that agents are able to get the recommendations from even those agents that are not known to it [13].

When any agent present in social network gets a request to provide recommendations about resource providers from his trustworthy acquaintance then it will recommend resource providers only if a. It has direct interaction with those service providers b. It gets the recommendations about them from its trustworthy acquaintances. The recommending agent will only keep those service providers in the recommendation list who are capable of doing job. The capability of any service provider is found out by computing the similarity between request vector as generated from query and expertise vector of service provider as generated from the attributes of services which it has already used. After finding the cosine similarity between the request vector and the expertise vector the recommending agent will decide upon the capability of any resource provider using the formula written in equation 1 as

nn

ii

n

iii

r

erERC

=

==⊗=

1

2

1

)(

(1) Where R (<r1….rrn>) is a request vector and E (<e1….en>) is an expertise vector which is prepared on the attributes of the services provided by service provider and n is the number of dimensions these vectors have. If C > Cmin , where Cmin is minimum value of the capability of the cloud providers desired to perform the job, then only the service providers’ will be considered for recommendation as direct experience. The recommending agent will also send this query to trustworthy acquaintances to get their recommendations if the Ttimeout permits where Ttimeout is the value till which the user can wait for the result of execution of the job.

Once all the recommendations are available with the agent, then the agent aggregates those recommendation on the basis of weights assigned to the direct and indirect information using the equation 2.

Agg. (Pi x) = W1 Pi x+ W2 (t1*Pi1 + t2*Pi2 + t3*Pi3 + …..+ t

n*Pin) (2)

Where Pix is the ith aggregated attribute of the service provider as done by the agent x and tk is the degree of trust of agent x on agent k and Pij is the aggregated ith attribute of service provider as send by agent j to agent x. Here W1 and W2 represents preferred weights assigned to the direct and indirect information by user agent x; here, W1, W2 € [0, 1] and W1 + W2 = 1.

C. Generating the recommendations for a user The fuzziness and uncertainty present in the recommendations collected from the feedback of trustworthy acquaintances is handled by Fuzzy Inference system i.e. (FIS). The advantage of using FIS is that it is able to narrow the gap between computational logic and human reasoning. Furthermore, fuzzy rules based on natural language reduce complexity and incorporate existing expert knowledge providing flexibility to the system. The steps to be followed for applying FIS is written below Algorithm: Fuzzy inference procedure to compute reputation

1. Find the membership function of every aggregated parameter i.e. Job Success rate and Job Computation rate.

2. Apply the fuzzy rule set to map those aggregated input parameters space to output reputation space.

3. Aggregate the fuzzy outputs of all the above rules written to a fuzzy set i.e. reputation.

4. Compute numeric crisp value of reputation with defuzzification of the fuzzy output.

In this framework, three triangular membership functions were taken for input and output. A rule set of forty have been taken for mapping input space to output space. After implication and aggregation, the defuzzification is done using the centroid of the trapezoid method using eq. 3

.

=

== n

ii

n

iii

P

PCreputation

1

1 (3)

where Ci represents centroid of the trapezoid obtained after aggregating the output of the fuzzy rules and Pi represents the firing strength of the ith fuzzy rule used to compute fuzzy variable fi using eq. 4 as

� �= =

=k

j

n

iiji dP

1 1

2

(4)

where Pi represents the firing strength of the rule Ri used to compute fuzzy variable fi and dij represents the degree of

714719716

membership of the input variable xi to the rule Ri. Here n represents the number of input variables and k represents the number of rules that gives output as fuzzy variable fi [3]. Once the reputation of all the recommended service providers are computed by the above method then the user agent prepares a list of service providers keeping in consideration the minimum reputation as specified by the user for job delegation. Then the user agent sends its applications to the service provider present in the list for its services

D. Updating trust level on trustworthy acquaintances The user agent gets the recommendations from various trustworthy agents and used them for the selection of service providers. Once the services of the service provider have been used, the user agent has to update the trust on recommenders. The updation is done on taking in account the difference in reputation values as obtained by aggregating the recommendations and as given by an agent [12]. Depending upon whether the difference is below a threshold dt, the user agent updates the trust value on recommenders as in eq. (5).

Trust(R) = Trust(R) + (dt- d) (5)

where d represents the difference between the aggregated value of recommendations at the user agent and the individual recommendation of the recommender. Hence if there is agreement, (dt–d) is positive then trust level on trustworthy acquaintance will increase but if there is difference in opinions then trust will decrease. This update process of the trust level is a continuous process and is done at every agent as and when the services of the service provider are used. It is stated that initially, it might be that some cloud providers may gain recommendations due to their brand but if they do not perform consistently in future then they will not get recommended by the trustworthy acquaintances and gradually they will get eliminated from the system whereas when once non performing entity starts performing then gradually it will gain the good reputation due to its recommenders. Thus the system will continue to evolve.

IV. Experimental Study Experimental study was conducted on the data of around 1GB collected from the site http://rtm.hep.ph.ic.ac.uk/ce.php for 31 days (11th December 2011 to 9th January 2012). It is 3D Real Time Monitor [10] project of GridPP which monitors thousands of nodes spread over more than 300 sites containing a total of approximately 633 computing entities joining in the EGEE Grid [7]. It provides the globe view using satellite imagery from NASA and displays running and scheduled jobs, job transfers and some of detailed information for each site on the Grid. All the information was kept in aggregated reports generated by that site about

each node present in the grid for each day. The snapshots of two such reports is shown in fig 1 It is taken as assumption that all these sites acted as different service providers providing services for storage and computational resources in the cloud. Report of antaeus.hpcc.ttu.edu

Report of atlasce01.na.infn.it

Fig 1: Reports of Two Computing Elements present in EGEE grid

A. Experimental results The framework was implemented on Intel (R) Core (TM)

i3-2330M CPU @ 2.20 GHz in Jade 4.0 on Net Beans 6.5. The data written in 17,690 reports (1.28 GB) were distributed randomly among different 25 agents as their knowledge base about service providers which will be used when asked for recommendations from trustworthy acquaintances. The web of trust existing between some of the agents is as shown in table 1

TABLE 1: WEB OF TRUST EXISTING BETWEEN USER AGENTS

Agents Trustworthy acquaintances

Trust level

Agent 1

Agent 2 0.5 Agent 3 0.7

Agent 2

Agent 3 0.7 Agent 4 0.8

Agent 3 Agent 4 0.8

Agent 4 Agent 5 0.7

Agent 5 Agent 4 0.7 Agent 6 0.2 Agent 7 0.4 Agent 8 0.3

Initially a random selection among 25 agents was done for user who had an application to be executed on the cloud with the specified requirements such as operating system (OS), server type, number of CPUs, size of memory required (GB),

715720717

CPU speed(MHz), storage size (GB). The user agent from those requirements generated a query as vector which it passed as a request to its trustworthy acquaintances for recommendations about service provider. The trustworthy acquaintances also passed the query to their acquaintances to get the recommendations if the time permits. Each of the acquaintances recommended aggregated attributes of the service providers using eq. 2 as shown in table 2.

TABLE 2: RECOMMENDATIONS AS PROVIDED BY TRUSTWORTHY ACQUAINTANCE

User Agent

Trustworthy

acquaintances

Recommended Service Providers

Job Success

rate

Job Computation rate

a5 a4 unipa-ce-1.pa.pi2s2.it 2.1 3.98 grid129.sinp.msu.ru 4.9 14.525

grid012.ct.infn.it 7.0 11.724 a6 f-ce02.grid.sinica.edu.tw 9.8 15.05

ce131.cern.ch 16.8 17.15 ce02.ncg.ingrid.pt 6.09 4.095

a7 ce02-cms.lip.pt 5.46 5.145 gridgate.scg.nuigalway.ie 1.4 18.25

a8 apcce01.in2p3.fr 1.26 2.94 ce002.ipp.acad.bg 5.67 4.830

ce.grid.eenet.ee 0.420 5.25

The attributes job success rate and job computation rate were taken from the %success and % useful computation attributes, respectively from the reports. For applying fuzzy inference engine, triangular membership functions were taken for input attributes which are shown in fig 2 as

Figure 2: Membership functions of Job Success rate, Job Computation rate and Reputation

A set of 42 fuzzy rules as shown in fig 3 which when applied with given aggregated parameters of job success rate and job computation rate then crisp value of the reputation of each service provider was obtained using FIS is shown in table 3.

Figure 3: A set of 42 rules applied to get repiutation

TABLE 3 : RECOMMENDED TRUSTWORTHY SERVICE PROVIDERS TO USER AGENT

Agent Service providers Reputation

a5 gridgate.scg.nuigalway.ie 8.016 ce131.cern.ch 7.14 ce02-cms.lip.pt 7.10 ce02.ncg.ingrid.pt 7.02 grid012.ct.infn.it 6.65 ce.grid.eenet.ee 3.92 unipa-ce-01.pa.pi2s2.it 3.843 f-ce02.grid.sinica.edu.tw 3.831 grid129.sinp.msu.ru 3.802 apcce01.in2p3.fr 1.21 ce002.ipp.acad.bg 0.22

The above list of recommended service providers is also shown in fig 4. Once the list of service providers is given to the user agent, user agent can decide among those recommended trustworthy service providers for its application and safeguard himself from loss of his sensitive data.

Fig 4: Recommended Service providers from Trustworthy Acquaintances Experiments were also performed by taking the recommendations from the random agents instead of

716721718

trustworthy acquaintances. Results of the experiments are shown in fig 5.

Fig 4: Recommended Service providers from Random Agents It is observed from the fig 4 and fig 5 that when recommendations are taken from trustworthy acquaintances then virtuous computing elements are recommended which shows that by employing the trust in the selection of service providers present in the cloud helps the user.

V. Conclusion In recent years, due to the immense growth of service providers in cloud computing environment, it has become increasingly difficult for the cloud user to find the service provider for its services because of the fear of loss of data. To handle this issue, a framework is proposed in this paper which helps the user in finding the trustworthy service providers in cloud environment using the recommendations given by trustworthy acquaintances. Experiments have been done on the real data considering various computing elements present in the grid as service providers of the cloud and a list of trustworthy service provider is being generated for the users to deploy his application on the cloud

References

[1]. A. Josang, R. Ismail, and C. Boyd, “A Survey of Trust and Reputation Systems for Online Service provision” Decision Support Systems, vol. 43(2), pp 618-644,2007.

[2]. B. Alunkal, I. Veljkovic, G. Laszewski, and K. Amin.,“Reputation Based Grid Resource Selection”, Proc. Adaptive Grid Middleware, 2003.

[3]. B. Gupta, H. Kaur, P. Bedi, “An Agent based Reputation System for Unreliable Grid Environment”, 3rd International Conference on Computer Modeling and Simulation(ICCMS 2011), pp 175-180, Vol. 1 Bombay, India, 2011

[4]. B.Gupta, H.Kaur, P.Bedi”, “A Reputation based Service provider Selection for Farmers”, Journal of information assurance and security (JIAS), 2010, pp. 515, Vol 5, ISSN: 1554-1010.

[5]. C. Everett, “Cloud Cter Fraudomputing- a question of trust”, Computer Fraud & Security, Vol, 2009, no 6, pp 5-7, 2009.

[6]. CSA, “security guidance for critical areas of focus in cloud computing v2.1,” Cloud Security Alliance, Tech, Rep.,2009

[7]. EGEE(Enabling Grids for E-sciencE) project. http://www.eu-egee.org/.

[8]. ENISA, “Cloud Computing, benefits, risks and recommendations for information security”, ENISA, tech. rep. 2009.

[9]. F.Azzedin,M.Maheswaran, “A trust brokering system and its application to resource management in public resource grid”, International Parallel and distributed Computing Symposium (IPDPS)2004.

[10]. GridPP. Real Time Monitor.http://gridportal.hep.ph.ic.ac.uk/rtm/

[11]. I. Mouline, “Why assumptions about cloud performance can be dangerous to your business,” Cloud Comp. J., vol. 2, no. 3, pp. 24–28, 2009

[12]. J. Abawajy, “Determining service trustworthiness in intercloud computing environments,” Int. Symposium on Parallel Architectures, Algorithms, and Networks, vol. 0, pp. 784–788,2009

[13]. P. Bedi and H. Kaur ,” Modelling Agent Reputation in Uncertain Environment”, Proc. International conference on Cognitive Science, Allahabad,2004, India.

[14]. P. Bedi and H. Kaur,”Trust Based Recommender System”, Proc. Int’l Conference on AI, 2005, pp 798-801, Las Vegas, USA.

[15]. P. Bedi, H. Kaur, and S. Marwaha. Trust based recommender system for semantic web. In Proc. of IJCAI’07, pages 2677–2682, 2007.

[16]. R. Buyya, C. S. Yeo, S. Venugopal, J. Broberg, and I. Brandic, “Cloud Computing and Emerging IT Platforms: Vision, Hype, and Reality for Delivering Computing as the 5th Utility”, Future Generation Computer Systems, Elsevier Science, 2009

[17]. S. Song, K. Hwang, and Y.K. Kwok. “Trusted grid computing with security binding and trust integration”. Journal of Grid Computing, 2005.

717722719


Recommended