多云工作流优化传输费用的数据布局策略

A Cost-Effective Data Layout Strategy for Multiple Scientific Cloud Workflows

科学工作流应用是一种复杂且数据密集型的应用,常应用于结构生物学、高能物理学和神经学等涉及分布式数据源的学科。数据分散存储在基于互联网的云计算平台上,致使科学工作流在执行时伴随着大量的数据传输。云计算是一种按使用量付费的模式,数据传输产生传输费用,尤其在多个工作流相互协同的情况下,将产生更高的传输成本。该文从全局的角度建立基于多工作流数据依赖图的传输成本模型,研究基于二进制粒子群算法（BPSO）的数据布局优化策略,从而减少对云计算传输资源的租赁费用。

Scientific workflow is a complex and data-intensive application.It often used in disciplines related to distributed data sources,such as structural biology,highenergy physics and neurology.Data distribute in Internet-based cloud computing platform,resulting in transferring mass of data by scientific workflow running.Because cloud computing is a pay-per-use model,so data transfer costs incurred.Especiallyin the case of multiplecooperative workflows,datatransmissionwill produce higher costs.Firstly,this paper based on multiple workflow data dependence graphestablish transmissioncost model.Secondly,this paperproposed anew particle swarm optimization-based strategy for cost-effective data layout in multiple scientific cloud workflows.The experimental results show that the strategy is much better than its traditional counterparts.