A Case for the Optimization of Database Workflow Orchestration in Heterogenous Environments

We present a novel optimization approach for the orchestration of the execution of database workflows in heterogeneous infrastructures, hereby specifically focusing on sorting algorithms. We give a case for the validity of our approach by developing a generic template for a family of optimization algorithms. We develop a model for the mathematical analysis of parallel sort algorithms and design a graph based algorithm for finding optimal complete binary work flow execution trees in networks aiming for maximum performance of the parallel Merge Sort. We justify our approach by an exhaustive simulation analysis, which leads to a refinement and optimization of the workflow orchestration process.