Trust Model In Cloud Computing

Published Date: 02 Nov 2017

ABSTRACT:

Cloud computing is the networking of virtual server which serves client request in ease of access manner. The main concern about cloud is thrust since many private and local data are shared and accessed in it. In this paper, a novel trust framework is proposed using Fuzzy Cognitive Mapping. The result of the application of this proposed framework yields thrust of each resources present in the system. These approximated thrust results of the system are propagated to the user who then decides the utilization of system.

INTRODUCTION:

Cloud computing is the advancement of grid computing which has been devised with ‘pay as per you use’ basis. This can also be said as on demand access of resources which reduces the installation of software or availability of hardware at the client side. All the resources can be accessed by the client in the virtualized manner. The cloud is established with the help of inter network of all virtualized server which can serve different client requirement [1].

The ways which cloud made available to the client are Infrastructure as Service (Iaas) in which entire infrastructure (including hardware and software) are made available to client, Platform as Service (Paas) in which platform (like OS) is made as service to client, Software as Service (Saas) through which any application software can be made available to the client and it can access remotely, Database as Service (DaaS) in which high end database can be used remotely by the client, Network as Service(NaaS) in which network is provided as services to the client when necessary.

The cloud services when access remotely by the client provides several advantages like easy access which does not requires installation at client side, consuming high level services, less overload at client side, documents made universal access, less investment[2].

The major concern about cloud computing is distributing resources to the client efficiently and establishing secure communication with client [3]. In order to overcome the concern in proper manner we have proposed thrust model in cloud services. Through thrust framework a thrust between cloud provider and consumer will be established and thereby cloud consumer can consume the cloud without any intervention. Fuzzy system is introduced in order to evaluate thrust framework mainly in IaaS.The result of evaluation of framework using fuzzy system will reveal the trustworthiness of the designed system.

The degree of thrust framework is measured using scalability, Availability, Security and Usability. These attributes are verified for the system which is designed and level of occurrence is checked and it ultimately leads to the result of degree of thrust of the system.

RELATED WORK:

Fuzzy-Cognitive Map is the logical based approach from which thrust of the system can be evaluated. It measures the impact of the system and analyse the relationship between several attributes provided for analysis. The system obtains the feedback from several customers like good, better or poor and stores it in the centralized database. The values like -1, 0, and 1 are assigned to different attribute. The feedback against those values is substituted and the resultant value is analysed.

Based on the user feedback the cloud thrust is measured [4] and supplied to the all the users who are new to the system. This system posses several disadvantages such as false feedback, duplication of users, competitors may intentionally giving improper feedback, etc. Also this model uses centralized database which causes single point of failure. When the centralized database data getting failed the entire system data gets crashes and thus leads to the destruction of entire system.

The revised system takes QOS of the system and the feedback from user to rank the resources [8].This system also uses peer-to-peer data storage in which the feedback details and the result of analysis about QOS of the system also stored in each system. This overcomes the drawback of previous system as the single point of failure cannot affect the entire system since redundant data are placed everywhere in the network of system. Monitoring agents are placed in the system which continuously monitors the system for the updating of the data placed in every system whenever the data in central database gets changed. The changes of the values can be easily propagated to the other system [7].

Monitoring agent also helps in avoiding cheating of the values by any user of the system. Each peer in the system provides feed back to the central data after collecting the information. The feedback given by every peer is different since the method adoption to calculate the feedback by every system is entirely different. These peers stores its feedback in UDDI registry services thus it can analyse and verify the web services trustworthiness [9].

The consumer’s infrastructure independent services are produced as the result in cloud computing. The methodology which has been proposed will yield a good metric for cloud providers. The effectiveness of the cloud also measured and supplied as the result to the user. The resultant yields the thrust of the cloud services which is more useful for the cloud requestors or the user especially who are new to the services.

APPROACH OF OUR RESEARCH:

In this model we are going to formulate the trust model for cloud computing based on the fuzzy approach. The various steps which we are going to perform while formulating the model are as follows:

Definition of the problem

Defining the variables of that domain

Data gathering

Problem Description

Domain Variable Definition

Gathering of data

Designing of the model

Implementation

Designing of the model

Implementation

Fig 1.Methodology of our approach

Problem Description: The existing infrastructure lacks in providing the trust based service selection in selecting the services offered by the service providers nowadays. The main issues which are not taken into consideration in the previous architecture are stated below:

(I) Prior to it no such trust model exists which could measure the trust of the cloud architecture. Yet model exists to measure the trust in online services but for cloud infrastructure it’s absent.

(ii) The providers of the cloud who provides services does not provide clear understanding of the factors of Service Level Agreement (SLA) contracts so that the users can sign the online agreement with the cloud providers.

(iii) The lack of the models to calculate the cost at each level of the cloud architecture.

In this paper we are going to calculate the trust of the cloud service providers so that the users can understand and utilize the services provided by the trusted cloud service providers efficiently.

Domain Variable Definition: During the Proposing the SLA agreement the various factors are taken to be in consideration while constructing the trust model for the cloud architecture. Depending upon the types of cloud like IAAS, PAAS, SAAS etc the impact factors for calculating the trust also varies. But in generality the following factors are identified for calculating the trust model .They are as follows:

(i)Scalability (Sca): It is termed as the ability to scale number of users i.e. the ability of the cloud service providers to provide the service seven if the number of users are increasing or scaling

(ii)Availability (Avi): It is the major issue in providing trustful services to the users. The ability of the service providers ensuring that their users will never go for a service outrage.

(iii)Security (Sec): Security is one of the major concerns which are required to be maintained by the cloud service providers to their users while providing services to them. Authentication mechanisms and various types of encryption schemes have been employed in order to achieve security in cloud computing area.

(iv)Usability (Usa): It is the measure of the services potential provided by the cloud service providers in order to accomplish the goals of the users.

Gathering of the data: As the approach which w e are going to design in this trust model based on the fuzzy approach so a lot of amount of dataset is to be gathered to which works as a training data. The data is collected by doing an online survey and asking a lot of question on the various impact factors selected for the construction of the trust model for the cloud architecture.

Designing of the Model:

We have used the fuzzy based logical approach to calculate the uncertainty in the in evaluating the trust calculation in the cloud environment. In this paper the we use three fuzzy sets for input factors and five fuzzy sets for the parameters of the output. The three input fuzzy sets are low (L), medium (M) and high (H).The five fuzzy sets for the output parameters are Very poor (VP), poor (P), good (G), Very good (VG) and excellent (E) .The set of sample values for the IAAS types of the cloud has been shown in the table 1.

Fig 2.Sugeno Fuzzy inference System

Table 1.Samples of Fuzzy rules for IAAS Cloud trust evaluation

The Basic steps for the trust evaluation model done in this paper have been shown by fig 3.which is depicted below.

Fig 3.The Model for evaluating trust in the cloud environment

Defining fuzzy sets and the inputs: The first step of this method is to set the value for each and every factor both for input and for output sets. The cloud service requestors can set the weights of the various factors on the basis of which the trust is going to be evaluated and if the value is not provided by the users then the system is setting the weights of the various factors as equal.

Fuzzification: Membership functions are used to assign the inputs the appropriate degree of the fuzzy sets.

Membership degree Calculation: The suitable membership is selected and the function determines how each of the values in the input range belongs to the input space of the membership value. The range of the membership function lies between 0 to 1.

Fuzzy rules Design: We need to propose the fuzzy rules for the inference engine, so that efficient services can be provided to the users on demand. Discussions and online surveys are made to calculate the fuzzy rules.

Inference Engine: We are using the Sugeno inference engine which is the most popular inference mechanism. The Sugeno inference engine takes the fuzzy sets of inputs and produces the outputs in the form of crisp sets.

Defuzzification: To produce a single output value we are using the Defuzzification approach which is using the centroid calculation.

CONCLUSION:

The application of proposed framework with FCM provides approximated thrust rate for the resources. This overcomes the existing thrust model since it relies on QOS and the user feedback of the system. This also avoids duplication of user and other illegal activities with the help of monitoring agents. Thus the proposed model results in the reliable thrust rate for several resources in the system which then helps cloud requestor to decide their utilization of that resource in the cloud system.