INTEROPERABLE FEATURES CLASSIFICATION TECHNIQUE FOR CLOUD BASED APPLICATION USING FUZZY SYSTEMS * C. Saravanakumar 1 and C. Arun 2 1 Department of Computer Science and Engineering, Sathyabama University, Chennai,Tamil nadu, India 2 Department of Electronics and Communication Engineering, R.M.K College of Engineering and Technology, Chennai, Tamil nadu, India *Author for Correspondence ABSTRACT Nowadays cloud computing plays a vital role in internet computing through the service oriented paradigm. The Cloud Service User (CSU) has requested the services from the Cloud Service Provider (CSP) for accessing the services which are deployed at the cloud environment. There are plenty of services available in the cloud which are provides the same serviceslike an image editing software. These software services will not provide a complete service to the requesting customer which leads the customer retention problem. The existing techniques are based on a single service provider who provides the service. Once the CSU selects the CSP for the service which are not gets full features required to process the request i.e. not complete solution. The interoperability has been achieved for satisfying the user for collaborating various service providers with complete solution. A Common Deployment Model (CDM) has been introduced with various functions for providing the QoS service to the customers. The proposed method focuses on the classification of the features based on fuzzy systems. This fuzzy classification method has implemented in the CDM which classifies the users along with their features for effective cloud interaction. The architecture of the proposed work provides a complete solution for the cloud based application. The Fuzzy Inference System has been implemented to classify the users with their access level. The member ship value is identified to classifythe user in three categories they are basic user, normal user and advanced user. In future this technique can be implemented the cloud security and privacy for protecting the customer s data. Keywords: CDM, Fuzzy systems, Interoperability, Cloud Computing and Fuzzy Classification INTRODUCTION Cloud computing grows enormously due the growth of internet and related development. The cloud computing is a service oriented computing which provides the service to the customer with on demand property. The origin of cloud computing is a desktop computing which is used only for developing a standalone application. The clientserver model has been introduced to share the data among the services followed by the distributed computing, collaborative computing etc., isalso a computing which comes before the cloud computing. There are various service models exists in the cloud computing namely SaaS (Software as a Service), PaaS (Platform as a Service), IaaS (Infrastructure as a Service). These services are deployed into the deployment models such as public cloud, private cloud, hybrid cloud and community cloud. These models are related to various parameters such as limit of access, capacity, security, reliability, trust and privacy of the cloud computing. Normally the CSU depends on a single CSP who cannot get a complete set of features for their current requirement. This problem leads the customer will not free to change the CSP, so the issue causes a customer retention problem. The one and only solution is interoperability which means that various CSP s can able give their services to enhancing the ability by establishing the Service Level Agreement (SLA) (Reddy et al., 2009). This is the contract among the CSP s who are not concentrates the different enhancement but they only target the revenues.there are many standards exits over interoperability of the cloud which are classified in to intercloud interoperability standards and internal cloud interoperability standards. The issues in these standards are solved by establishing the transport mechanism among the virtual machine (Lewis, 2012) and to develop an API s for creating an interface with IaaS cloud (OFG, 2013) and to standardize the cloud storage interactions (Kresimir & Zeljko, 2010). The CDM has been introduced with various functions namely security management, service management, location awarenessover cloud computing, traffic shaping and SLA management (Saravanakumar & Arun, 2011, 2012, 2013). These features are used to standardize the cloud services in order to achieve maximum retention of the customer with high quality. Fuzzy keyword search overcomes the problem of traditional plain keyword search to protect the cloud data in an accurate manner. This technique is not suitable for complex natural language and multiple semantics over the secured data (Syeda & Rangaraju, 2013). The fuzzy keyword searching with safe index is proposed for searching data in large dataset by using bloom filter. The size of the fuzzy set is large so the efficiency will be reduced so this will be carried out carefully to achieve quality response (Lixi et al., 2013). The fuzzy logic over cloud computing gives an 84
interoperability among various organizations for accessing a huge amount of data and also improves the accessibility of mobile cloud. This system is not suitable for precise and imprecise data (Rinku & Jasutkar, 2012). The fuzzy keyword search displays a similarity keyword and also achieves maximum security and preserve the cloud data. It is not suitable for conjunction of keywords and sequence of keywords to achieve relevancy in the search result (Ranjeeth Kumar & Vasumathi, 2011). The challenging issue of cloud computing is solved by introducing a novel load balancing algorithm with fuzzy logic over various cloud elements. The Fuzzy Based Round Robin load balancing algorithm compares and proves better result than the Round Robin load algorithm (Sethi et al., 2012).The K-grams based fuzzy keyword search has been introduced to solve the issues like spelling errors and morphological variant used in the searching process (Wei et al., 2013). The existing work fully based on the searching of data which are hosted in the cloud using keywords selected by the customer. These works are not addressed in the classification of data and the customer with their access level. The proposed work focuses on the fuzzy based classification and prediction of application features and it selects a suitable service for the customer. The customer doesn t know about the quality of the services which comes from the cloud computing end. The request of the cloud comes from various sources i.e customers who need the services. The proposed work classifies the user according to their access level namely basic user, normal user and advanced user. The basic user request only minimum features supported for the cloud request. The normal user requests the service with some conditions whereas the advanced user needs only maximum features supported by the application. The features are extracted from the cloud and classify the features using the fuzzy systems. These classified features are then selected for further prediction. The main objective of the proposed work is to select the suitable services in order to achieve a high quality service response to the customer. The paper is organized as follows; Section 1 represents the overview of cloud computing and related concepts. Section 2 describes the system of fuzzy based feature classification. Section3 represents the architecture diagram of the proposed system. Section 4 represents the implementation of fuzzy inference system. Section 5 proposed the method of the cloud service support using fuzzy system. Finally, Section 6 represents conclusion and future work. FUZZY BASED FEATURE CLASSIFICATION SYSTEM Fig. 1 shows that the fuzzy based feature classification system. In the cloud computing concept there are various application gives the same service. Suppose the customer needs the software to edit an image will generate the requests to the cloud. The customer should know about the application quality and accuracy. This parameter mainly depends up on the response from the CSP which gives the application with maximum features. Fuzzy based Feature Classification Application Selection Feature Selection Feature Prediction Feature Classification Feature Extraction Figure 1: Fuzzy based Feature Classification System The application selection is the first and foremost step in the proposed work. The features are extracted from the selected cloud application. The extracted features need to be assigning the membership value for classification. These membership values are identified on based the user access level. The features are compared with the features which are identified during the cloud request. The features are predicted for an efficient cloud response and also for customer retention. 85
ARCHITECTURE OF THE PROPOSED SYSTEM The CSU requests a service from the CSP by generating the cloud request which are handled by the deployment controller. The CDM is available in the deployment controller which provides various functions such as management of security, traffic shaping, service management, location awareness of cloud storage etc. The proposed method focuses on the fuzzy based feature classification are also included in the CDM. The main objective of the proposed work is to classify the users based on the strength of the access level. Normally the user has classified as a basic user, normal user and advanced user. These classifications are based on the features which are extracted from the cloud application. The extracted features are identified as a separate group based on their member ship values. The membership value of the basic user is identified with the minimum number of features selected by the user. The member ship value of the advanced user is identified with maximum number of features which are selected by the user whereas the normal users have some set of features. Figure 2: Architecture of the proposed system Fig. 2 describes the architecture diagram of the proposed work. The flow of the work is organized are as follows 1) the CSU request the cloud service 2) the credentials required for the cloud interaction is requested 3) credentials are send to the deployment controller 4) fuzzy based classification are carried out 5) the cloud response are send to the CSU. Fuzzy based feature classification method first selects the applications which are requested by the CSU and the features are extracted from the cloud storage. The extracted features are classified based on the fuzzy membership value and then the features are predicted. The prediction is used to assess the relevancy of the cloud application in order to achieve maximum customer retention. IMPLEMENTATION USING FUZZY INFERENCE SYSTEM Most real-world problems are represented as logic to process incomplete, imprecise, vague or uncertain information. If two valued logic is not suitable for real time environments, the fuzzy logic and fuzzy sets may give formal tools for uncertain information. The certainty of an element over the set can be represented using the membership function with related variables. A membership function is referred to the characteristic function of the fuzzy set. The Membership functions of fuzzy sets are represented in different shapes with the range of {0, 1}. In fuzzy system the dynamic behavior of the system can be categorized by fuzzy rules. Fuzzy rules are simple if-then rules i.e., the condition part referred as ancedents where as the conclusion part is called as consequent. The rules are applied into the FIS (Fuzzy Inference System) for further processing of fuzzy sets. FIS consists of fuzzification process, inference engine, defuzzification process, knowledge base, input and output sets. There are two FIS are most commonly used namely Mamdani and Takagi-Sugeno fuzzy Controller. Mamdani fuzzy controller is used in this implementation. A Centroid method of defuzzification process can be used for converting a fuzzy result set into 86
fuzzy output (Engelbrecht, 2007). The cloud application features are classified and predicted using the fuzzy system with related logic. An application features are classified into three main classes of the users based on the level of accessibility on cloud based word application they are basic user, normal user and advanced user. The basic user can able to use the cloud based word processing application with only basic features like creating, editing, printing etc., The Normal user can able to use the cloud based word processing services with basic features with some extra features like sharing, portability etc., An advanced user can use the full features in the cloud application. The goal of these classifications is to improve the quality of services for the cloud customer from the unnecessary ambiguous in service selection. The following work describes an implementation result of the cloud interoperability over cloud based word processing application. CLOUD SERVICE SUPPORT USING FUZZY SYSTEM The actual service supports the cloud based applications are shown in Fig. 3 which lacksin the usage application features in various cloud users. The membership value is calculated based on the fuzzy rules fixed using the MATLAB software. The lower and upper bound of the membership value is fixed based on the fuzzy rules. The membership value of the basic user is identified based on the features selected by the user. The range of service support is from 0.14 to 0.29 gives minimum feature count to the current cloud request. The lower bound value of the entire application request of various users is starts from 0.14 to 0.71. The user category falls into the given range based on the features selected by the users. The following membership functions are identified based on the values of the upper and lower bound of the service support. The membership functions of cloud service can be specified are as follows,... (1) (2) (3) (4) (5) Figure 3: Service Support map on cloud based word processing application 87
The features will not support while it exceeds the value 0.71 i.e. the result leads to cloud features unavailability. According to the result, the service will not support to the Cloud Service Users at certain limit, so the cloud services offered by the CSP to CSU is possible with interoperability. CONCLUSION AND FUTURE WORK Cloud computing provide the services to the customers for getting an effective and efficient access of their requirement. There are various services available in the cloud computing which are related to the CSP. The customer depends upon a single CSP who never gets complete services because not all the CSP provides the solution. The interoperability is the main solution for collaborating various CSP who provides a complete solution. CDM gives the solution for interoperability with various features which also deployed over deployment manger. The classification of the users with access level is important then only the level of interoperability is fixed. The proposed work focused on the classification of the users and features using the fuzzy systems. The membership value is calculated and then identified by the user with their access level. The fuzzy classification is used to select the application, extract the features, select the features and predict the features. These features provide a reliable service to the user in order to retain the customers. The implementation of interoperability can be achieved by using the fuzzy inference system with fuzzy rules. The fuzzy based classification method will be extended to the cloud security and privacy with high reliability. REFERENCES Engelbrecht PA (2007). Computational Intelligence: An Introduction. 2 nd Edition, University of Pretoria, South Africa 451-478. Kresimir P & Zeljko H (2010). Cloud Computing Security Issues and Challenges. MIPRO, May 24-28,Opatija, Croatia 344-349. Lewis GA (2012).The Role of Standards in Cloud Computing Interoperability. TECHNICAL NOTE,CMU/SEI- 2012-TN-012,Research Technology, and System Solutions Program, Carnegie Mellon University, Available at: http://www.sei.cmu.edu,http://www.sei.cmu.edu/reports/12tn012.pdf(accessed 14 August 2013). Lixi Liu, Chi Zhang, Shaowen Yao, Shipu Wang & Wei Zhou (2013). Fuzzy Keyword Search with Safe Index over Encrypted Cloud Computing. TELKOMNIKA, 11(10) 5884-5889. OGF (2013). About Open Grid Forum. Available at: http://www.ogf.org/about/abt_overview.php (Accessed 14 August 2013). RanjeethKumar M & Vasumathi D (2011). A novel implementation of fuzzy keyword search over encrypted data in cloud computing. International Journal of Computer Trends and Technology, 275-279. Reddy BK, Ramacrishna Paturi V & Atanu R (2009). Cloud Security Issues. IEEE International Conference on Services Computing 517-520. Rinku R& Jasutkar RW (2012). Fuzzy approach to mobile cloud computing. International Journal of Recent Technology and Engineering, 1(2) 89-91. Saravanakumar C & Arun C (2012). Secure Framework for publishing and accessing the cloud services with common deployment. European Journal of Scientific Research, 1-8. Saravanakumar C & Arun C (2012). Traffic analysis and shaping of the cloud services over common deployment model using cloud analyst. International Journal of Computer Applications, 43(4) 33-35. Saravanakumar C & Arun C (2012). Instance management for software application of the cloud over common deployment model. Emerging Trends in Science, Engineering and Technology 63-67. Saravanakumar C & Arun C (2013). Location awareness of the cloud storage with trust management using common deployment model. ICCCNT IEEE. Saravanakumar C & Arun C (2011).Interaction among Cloud Services on Common Software Application with Service Level Agreement. International Journal of Electronics Communication and Computer Engineering, 2(2) 1-4. Sethi S, Sahu A & Jena SK (2012). Efficient load balancing in cloud computing using fuzzy logic. IOSR Journal of Engineering, 2(7) 65-71. Syeda Farha Shazmeen & Rangaraju D (2013). Using different searching schemas for fuzzy keyword search over cloud data. Graduate Research in Engineering and Technology: An International Journal, 1(2) 41-44. Wei Zhou, Lixi Liu, He Jing, Chi Zhang, Shaowen Yao & Shipu Wang (2013). K-Gram based fuzzy keyword search over encrypted cloud computing. Journal of Software Engineering and Applications, 6 29-32. 88