Forecasting software aging of service-oriented application server based on wavelet network with adaptive genetic algorithm

Abstract

Web services are gaining acceptance as a standards-based approach for integrating loosely coupled services. Achieving high levels of reliability and availability of service-oriented application server in spite of service or infrastructure failures poses new challenges. According to the characteristic of performance parameters of service-oriented application sever, a new software aging forecasting model based on wavelet network is proposed. The dimensionality of input variables is reduced by principal component analysis, and the structure and parameters of wavelet network are optimized with genetic algorithm and evolutionary programming. The objective is to observe and model the existing systematic parameter data series of service-oriented application server to forecast accurately future unknown data values. By the model, we can get the aging threshold before application server fails and rejuvenate the application server in autonomic ways before observed systematic parameter value reaches the threshold. The experiments are carried out to validate the efficiency of the proposed model and show that the aging forecasting model based on wavelet network with adaptive genetic algorithm is superior to the BP neural network model and wavelet network model in the aspects of convergence rate and forecasting precision. © 2007 IEEE.