A Fixed-Structure Learning Automation Solution to the Quality Aware Application Service Configuration in a Grid Environment

Sharath Babu Musunoori, and Geir HornSimula Research Laboratory1325 Lysaker, Norwayemail: f sharath, geirhog @simula.noABSTRACTThe significant advancements in grid computing providea promising platform for parallel and distributed systems.However, achieving acceptable application performance inthe grid environment remains a difficult engineering chal-lenge. In particular, a class of applications that requires thefulfilment of some quality requirements in order to satisfythe user needs. In this paper we introduce an algorithm toallocate an application’s services to computing nodes. Theproposed algorithm is the very first designed and imple-mented based on the concepts of learning automata. Thealgorithm has been rigorously tested and evaluated throughextensive simulations on randomly generated applicationservices. The experimental results clearly show that thelearning automata based approach is a promising solutiontechnique for configuring applications based on composingservices in the grid environment.KEY WORDSService Configuration; Scheduling; Learning Automata;Grid Computing.