Open AccessA Model based Approach for Evaluating the Reliability of the Services in the Service-oriented Architecture
Author(s) -
Behzad Alipour,
Ali Haroon Abadi
Publication year - 2016
Publication title -
international journal of software engineerung and technologies (ijset)
Language(s) - English
Resource type - Journals
ISSN - 2302-4038
DOI - 10.11591/ijset.v1i1.4562
Subject(s) - computer science , reliability (semiconductor) , reliability engineering , certification , markov chain , service oriented architecture , service (business) , architecture , markov model , software quality , variance (accounting) , frame (networking) , software architecture , fault tolerance , distributed computing , software engineering , software , web service , engineering , software development , computer network , machine learning , operating system , art , economy , law , business , visual arts , power (physics) , accounting , quantum mechanics , political science , physics , economics , world wide web
Service-oriented architecture presents a frame in which the system functions are defined as a series of the distributed services in the intended sizes of the organization. These services are called by the other software and also are used for building the new services. Although this architecture offers a simple solution for building the distributed systems with loosely coupling, it introduces some additional concerns. One of the main concerns in designing a SOA system is general reliability of the system. Then the new technique for modeling the reliability is needed for certificating the services. Regarding to this weakness, in this paper, a certification method for the reliability in which services have been simulated as the discrete Markov chains, this work presents a model for estimating the reliability by exploiting the mean and variance of the visits obtained from analyzing Markov chain and integrating them into the reliability of the characteristics of each individual service. Results show that in the proposed method, less fault-tolerance than the recent methods for predicting the reliability of the systems is used.