A Combinatorial Optimization Algorithm for Load Balancing in Cloud Infrastructure

摘要

Cloud Technology is considered as the next generation of Business Process Outsourcing (BPO). It is a fast-evolving service oriented business which is based on Pay-As-You-Go model. Business analysts say that it is a double-edged sword service trade between cloud users and cloud providers. While cloud users are demanding to minimize turnaround time (response time), cloud providers bargain to maximize the profit by optimizing the utility rate of cloud resources. It ends up in a cost-benefit trade-off between the cloud providers and cloud users. In this paper, we propose a Load Balancing (LB) model that optimizes the cost-benefit trade-off between the cloud providers and cloud users. We first formulated the load balancing problem in Cloud as a combinatorial optimization problem. Subsequently, we designed a Modified Branch and Bound-Based Knapsack (MB3K) algorithm to solve the combinatorial optimization load balancing problem. The proposed model eases the cost-benefit trade-off significantly between the cloud users and providers. It is also empirically illustrated that the proposed algorithm performs better than the branch and bound knapsack algorithm and round robin scheduling algorithm.