Joint Customer/Provider Evolutionary Multi-objective Utility Maximization in Cloud Data Center Networks

Abstract

Cloud computing has found an important role in cost-effective and elastic network resource allocation community. Most researchers and practitioners have focused on cloud provider side for efficient resource allocation to end-users and to warrant their service-level agreements (SLAs). Due to heterogeneous user demands which request services from cloud data center networks, simultaneous satisfaction of all users to their pre-defined SLA level while making the cloud provider and all of the data centers to be cost-effective in terms of the financial income and energy-efficiency is a challenging issue. To address this problem, in the current work, the efficient joint provider/customer efficient resource allocation has been first formulated as a multi-objective optimization problem with convex constraint set in terms of virtual machine (VM) parameters (e.g., storage, CPU core, memory, etc.). Then, an evolutionary solution based on cooperative co-evolution algorithm has been proposed that can conduct a divide-and-conquer methodology for solving the complex optimization problem. Numerical results show the prominent feature of the method in terms of VM usage effectiveness and overall joint network cost/user SLA satisfaction.