ARCHITECTURE FOR DETECTING INFINITE LOOPS OF WEB SERVICES USING TIME BOUNDARY VALUES

Abstract

A web service is a remote system, implemented on a distributed system with request and response. To guarantee good services, software developers must find techniques for detecting and preventing errors during run-time. One problem that is important to such services is the infinite-loop situation, since the system will not return any results to the requester. Moreover, the resources of the system are overloaded until the server cannot function. Therefore, this research proposes the web service architecture relating to non-deterministic loop algorithms. For solving the infinite-loop problem, this research proposes two techniques: checking execution time by defining the execution time boundary, and checking patterns of iteration variable values. The techniques can detect and control the infinite-loop situation before causing damage to the system. The experiments using the proposed architecture measured the CPU usage. The results show that the architecture can improve the CPU usage. Moreover, the infinite-loop situation can be detected and terminated by the mechanisms of this architecture.