The Apache Airavata framework manages remote scientific application executions on a wide range of distributed resources, including clouds and supercomputers.  Airavata applications focus on workflow execution patterns such as directed acyclic graphs (DAG) with loops, conditionals, and other interrupts. An important challenge for Airavata is to move beyond traditional scientific computing workflows to challenges in data-centric (“big data”) computing and analytics with special emphasis to avoid migration of observational, experimental and/or simulation data among workflow processes. In this talk, we present an investigation of two case studies: the use of Airavata to power Apache Mahout’s execution patterns, and the integration of Apache Spark as a backend resource for workflows. Airavata community has access to academic clusters and will work as perfect test bed for big data solutions.