Multimodal pattern-oriented software architecture for self-configuration and self-healing in autonomic computing systems

Abstract

Because of the diverse nature of software systems, it is unlikely that systems will be developed using a purely service or component programming paradigms. Therefore, the ability to combine the strength of various programming paradigms and use them in a complementary manner becomes essential. As far as we know, there are no studies on composition of design patterns and pattern languages for autonomic computing domain. The work presented in the paper is concerned with composition of existing design patterns which are taken from various programming paradigms that are used for developing of the self-configuration and self-healing characteristics of the autonomic computing systems. In this paper we propose multimodal pattern-oriented software architecture with composition of Worker Object, Look-Up, Row Data Gateway Database access, Adaptation Detector, Case-Based Reasoning, Leader/Followers, and Observer design patterns using Java Web Services (JWS), JUDDI service repository, and Java Aspect Components (JAC) Frame work by Providing multimodality among the application working modes at various levels. We have also focused on the Data-Mining Association Rule Based Learning concept to introduce new service as composition of two or more services and thereby reducing the number of client requests to handle.