Consider The Virtual Machine Life Cycle And The Delay Of The Cloud Data Center Energy Saving Algorithm Research

In the Era of Cloud computing, efficient and economic management of resourcesin data centers is a challenging problem. Data centers consume a lot of power, andcause environmental pollution. Therefore design of energy-efficient algorithms tomanage data center resources and reduce operating costs has great significance.In order to reduce energy consumption of data centers, a large number ofexperiments are carried on physical servers. Analytical models of energy consumptionfor data centers are obtained by analyzing the data. This thesis proposes an offlineenergy-saving scheduling algorithm with delay and an on-line energy-savingscheduling algorithm with both delay and migration. Main features of two algorithmsinclude consideration of the life cycle of a virtual machine, two different scenarios ofthe offline and online, migration and delay factors in the scheduling process. Throughefficient allocation and migration of virtual machine requests the proposed algorithmscan minimize the total number and total running time of physical machines in the datacenters, and reduce the total power consumption in the long-run.Finally, the two proposed energy-saving scheduling algorithms are compared toother four existing energy-saving scheduling algorithms. In terms of the total energyconsumption of data centers, the total running-time of all physical machines and thetotal rejected number of requests, the proposed algorithms have distinct advantages.For the same set of requests, the two energy-saving scheduling algorithms have lesstotal energy consumption in the data center, less number of total physical machineused, and less total rejected number of requests.