DDS: integrating data analytics transformations in task-based workflows

High-performance data analytics (HPDA) is a current trend in e-science research that aims to integrate traditional HPC with recent analytic frameworks. Most of the work done this field has focused on improving frameworks by implementing their engines top technologies such as Message Passing Interface. However, there lack integration from an application development perspective. workflows have own parallel programming models, while (DA) algorithms are mainly implemented using transformations and executed like Spark. Task-based models (TBPMs) very efficient approach for workflows. Data can also be decomposed set tasks task-based model. In paper, we present methodology develop HPDA applications TBPMs allow developers combine seamlessly. A prototype been PyCOMPSs model validate two aspects: seamlessly developed better performance than We compare our results different programs. Finally, conclude idea integrating DA into evaluation method against Spark.