多云架构下成本感知的科学工作流中间数据布局策略

A Cost-aware Intermediate Datasets Placement Strategy for Scientific Workflow in multi-Cloud

首先构建了多云混合云环境,通过分析不同服务提供商计费策略的差异性和网络连接情况,建立相应的中间数据布局成本模型和传输时间模型,进而提出了基于离散粒子群的中间数据布局策略,以成本和时间为目标优化布局方案。仿真实验表明,相比于聚类算法,文章提出的策略能够显著降低成本。

This paper first builds the multi-cloud environment. Then, we build the cost model and the time model by analyzing different cloud service providers＇ price policy and network connection. Furthermore, we proposes the placement strategy based on discrete particle swarm optimization algorithm, aiming at optimize the cost and time. Simulation results confirmed that our placement strategy can significantly reduce the cost compared with the clustering placement strategy.