Reliability Prediction Model for SOA Using Hidden Markov Model

Reliability prediction models for Service Oriented Architecture (SOA) based applications have thoroughly discussed system level & component level details. However, significant difference between applications based on service-oriented architecture and traditional distributed application is marshaling of remote web services, nonetheless, reliability of these web services have not been discussed significantly. Recently developed strategies did not discuss how we can establish a mechanism which helps to predict reliability of internal structure of web service that may have involved either in failure or slow response. In this paper, we have specifically focused on the area of predicting reliability parameters like response time & throughput of hidden states that exist in any remote web service. Our proposed model will help end users to predict probability of success or failure of underlying hidden states of web services. Along with reliability it shall further help us to compute probability of scalability of those hidden states.