Application performance prediction in autonomic systems
Shobhana Kirtane, Jim C. Martin · 2006
An autonomic system is an intelligent system that is capable of self-configuration, self-healing, and self-management. Application performance prediction is a powerful tool that can be used in an autonomic system. Predicting application performance based on current or anticipated conditions provides fine-grained information that increases the chances that the autonomic manager makes correct decisions. In this paper, we report on the design and implementation of a system that can be used by an autonomic manager to predict the response times of transaction-oriented applications. Preliminary results suggest that our method leads to an average prediction error of less than 15% over a range of network and server loads.