When orchestrating highly distributed and data-intensive Web serviceworkflows the geographical placement of the orchestration engine can greatlyaffect the overall performance of a workflow. Orchestration engines aretypically run from within an organisations' network, and may have to transferdata across long geographical distances, which in turn increases execution timeand degrades the overall performance of a workflow. In this paper we presentCloudForecast: a Web service framework and analysis tool which given a workflowspecification, computes the optimal Amazon EC2 Cloud region to automaticallydeploy the orchestration engine and execute the workflow. We use geographicaldistance of the workflow, network latency and HTTP round-trip time betweenAmazon Cloud regions and the workflow nodes to find a ranking of Cloud regions.This combined set of simple metrics effectively predicts where the workfloworchestration engine should be deployed in order to reduce overall executiontime. We evaluate our approach by executing randomly generated data-intensiveworkflows deployed on the PlanetLab platform in order to rank Amazon EC2 Cloudregions. Our experimental results show that our proposed optimisation strategy,depending on the particular workflow, can speed up execution time on average by82.25% compared to local execution. We also show that the standard deviation ofexecution time is reduced by an average of almost 65% using the optimisationstrategy.
展开▼