Workflows have evolved as the natural tool for scientists to model their eScience experiments. With the scientific world producing data at an explosive rate, workflows have an important part to play in the end to end management of scientific data. To illustrate, workflow can help with fault tolerance and ease of administration when ingesting massive quantities of data using commodity hardware. The ability for workflows to automatically collect provenance on derived scientific data improves data discovery and publication capabilities. With better support for interoperating with data centric tools, workflows can become ubiquitous systems for scientific collaboration.