Performance evaluation of parallel policies based on system workflow in public cloud

Performance evaluation of parallel policies based on system workflow in public cloud

Pan Li Shuping Wang Ziping dragon Dan

This paper analyzes the parallel processing performance of phylogenetic genome workflow, and proposes a performance evaluation workflow for sciphylomics execution in cloud computing platform. Firstly, the application of mapping simplification model is introduced to realize Hadoop; then, the Scicumulus Cloud Workflow engine is presented, and finally, two parallel execution methods (Scicumulus and Hadoop) are implemented on the Amazon EC2 Cloud. The experimental results show that although the phylogenetic genomics experiment is strict to the computational environment, the experiment is still suitable for the implementation in the cloud. In addition, the evaluated workflow presents many characteristics of a set of data-intensive workflows that can be extended to other experimental types.

Performance evaluation of parallel policies based on system workflow in public cloud