Research On Data Placement Strategy For Green Cloud Computing

Cloud computing provides abundant storage and processing resources for the growing data information.How to design an effective data placement method to ensure the resources optimization and environmental protection of cloud data center has become a key issue to promote the healthy and steady development of the green cloud computing industry.Researches on data placement for green cloud computing mainly focus on isomorphic tasks and heterogeneous workflows.Isomorphic task-oriented data placement focuses on the impact of massive and complex data on the energy consumption,while ignoring the trade-off between efficiency and data collision.Heterogeneous workflow-oriented data placement focuses on the dependency relationship between data and tasks,while ignoring the impact of concurrency of tasks,multi-source,collision and sharing of data on workflow execution and energy consumption.Therefore,data placement for green cloud computing faces following challenges: 1)Data placement for isomorphic tasks ignores the trade-off between energy consumption and execution efficiency;2)Data placement for heterogeneous workflows needs to consider the impact of data multi-source,sensitivity,sharing and task concurrency on resource utilization and energy consumption,meanwhile ensuring workflow execution efficiency.In view of these,this paper conducts research on the data placement for green cloud computing.The main works are as follows:(1)For considering the energy consumption and performance of the cloud data center comprehensively,the energy-efficient and performance-optimized data placement method for isomorphic tasks is proposed.Technically,the data access time model and energy consumption model are established firstly.Then,sets of optimal placement strategies are obtained based on the Non-dominated Sorting Genetic Algorithm II(NSGA-II).Finally,based on the Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS)and Multiple Criteria Decision Making(MDCM),the placement strategy with optimal utility values is achieved to optimize the energy consumption and execution efficiency.(2)According to the characteristics of multi-source,sensitivity and sharing of data resources in heterogeneous workflows,a multi-objective optimization data placement method for heterogeneous workflows supporting green cloud computing is proposed.Technically,the resource usage model,access time model,energy consumption model and data sensitive conflict model for heterogeneous workflows are established firstly.Then,sets of optimal placement strategies are obtained based on the Non-dominated Sorting Genetic Algorithm III(NSGA-III).Finally,based on the TOPSIS and MDCM,the placement strategy with optimal utility values is achieved to optimize the resource utilization,data access time and energy consumption of cloud data center,meanwhile meeting the sensitive preservation and workflow deadline.