Performance Predictive Model of ESB as Service Integration Grows: The Case of ethio telecom

Service Oriented Architecture (SOA) is an architectural pattern that is based on the concept of a service and addresses, the requirements of loosely coupled, and protocol independent distributed computing. SOA used enterprise Service Bus (ESB) to realize its principles. ESB is an architecture to overcome the limitation of traditional architectures. One of its main functions is message transformation, which is the conversion process of data structure among heterogeneous systems during integration. The second function is routing functionality between runtime services which may introduce performance overheads. Ethiotelecom has implemented ESB to integrate businesscritical systems , due to additional service integrations, ESB's performance problem becomes a challenge to meet service level agreements. The main objective of this thesis is to develop and implement a performance predictive model, which can be used to evaluate the performance of the ESB system. Queueing Petri nets are powerful formalism that can be exploited for modeling distributed systems and analyzing their performance and scalability. By combining the modeling power and expressiveness of queueing networks and stochastic Petri nets, queueing Petri nets provide a number of advantages. The model can be useful for service quality and performance management process during design and service operation times. The Jmeter measured performance data is used for comparison and validity of the Queuing petri net network model of the case study system. Both in the experiment and the model similar performance impacts were observed for variety of workload inputs. Clearly one can see the level of the impact for each performance metrics as integration grows. The validation result indicates, the practicality of the model to do performance prediction and capacity estimation during system design and service operation of integrated applications in a heterogeneous environment. Moreover, unlike the existing system in which the case study is based upon, the model solution is further applicable for new capacity demands in similar system planning works.