Trust based resource selection in grids using immune system inspired model

Abstract

Grid is emerging as a potential technology leading to the utilization of the computer resources that are underutilized. Resources scheduling is one of the core grid services and has a direct impact on the overall performance of the system and the service quality. Scheduling in grids helps to find an better resource which improves the performance and also helps in utilizing the available resources efficiently. Most scheduling and resource selection algorithms do not take the characteristics of the resource into account. This can ultimately lead to reliability and security issues which will affect the quality of service. In this work it is proposed to implement a trust module which predicts trust across the grid network drawing principles from the Immune System (IS). By integrating the trust module with dynamic scheduling, security is increased and ratio of job task failure decreases. The proposed immune inspired system improves the makespan by almost 35% compared to traditional mechanism of manually selecting nodes based on trust and reputation.