الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 95 |  |  |
| | | from | 95 | | |
Abstract
Cloud Computing created a new era of network design, where users can get their required services without having to purchase expensive datacenter infrastructure for operating the environment, and eliminate hardware troubleshooting. These benefits created new business opportunities for small, medium, or even large size companies. One of the drawbacks of the cloud is the high power consumption, which becomes a challenge facing the Cloud’s Service Providers (CSP) to operate their huge datacenters. The most efficient solution for this challenge is Virtual Machine (VM) migration; VM migration is a technique that migrates VMs from under-utilized to other nodes, and switch under-utilized nodes to sleep mode, which saves the nodes’ power, but over-utilized nodes result in Service Level Agreement (SLA) violation. According to the work in this thesis, a VM migration with more than one policy is introduced, by dividing the CSP’s datacenter into many clusters. By applying only one policy may not be suitable for most of the cloud’s lessee, as some customers need critical SLA requirements regardless the cost, and others care about the leasing price. First we propose an adaptive power panel, which allow the CSP to divide the datacenter into clusters, where different policies can be applied on each cluster independently. The proposed adaptive power panel contains three classes, which are power aware, balanced, and high performance mode. Power consumption and SLA are adjusted based on the expected market needs. As an extension to that power panel, we also propose an adaptive technique using fuzzy logic which logically migrates a particular node from a cluster to another cluster based on the current workload.