OPTIMIZING DYNAMIC PROVISIONING OF RESOURCES IN MULTI-TIER CLOUD COMPUTING

Abstract

In order to meet Service Level Agreement (SLA) requirements, optimal resources in Cloud computing need to be provisioned with minimum execution time. Most existing researches on resource provisioning consume considerable execution time because it provides resources using Single-tier Clouds. Our research innovatively applies five meta-heuristic algorithms for dynamic provisioning of resource in Multi-tier Clouds, with the purpose of reducing total execution time. The algorithms adopted are Simulated Annealing (SA), Particle Swarm Optimization (PSO), Simulated Annealing-Particle Swarm Optimization (PSO-SA) hybrid, Ant Colony Optimization (ACO), and Artificial Bee Colony (ABC). This thesis also presents a comparative study of the results of the suggested methods against results of the same algorithms in Single-tier Clouds to validate our findings. Simulation results of the proposed algorithms show an improvement in resource provisioning by adopting meta-heuristics into Multi-tier Clouds, which proves to perform better than in Single-tier Clouds. Specifically, ACO algorithm was recorded to require less execution time for resources provisioning than the other algorithms in a Multi-tier Clouds. Moreover, ACO algorithms can reduce execution time by as much as 12.7% compared with execution time of ACO algorithms in Single-tier.