On the appropriate granularity of activities in a scientific workflow applied to an optimization problem

Scientific workflow management is an active area of research and development responding to an increase in the complexity of computational models, data analysis and data size. In the authors' experience, the distinction between scientific workflow software (SWS) and modelling frameworks or toolsets is at best not clear in the minds of users or developers, at least in the hydrology domain where the interest in concept of scientific workflow appears rather recent. It is understandably tempting for some users to assume that these new software tools aim to replace their existing modelling tools, with varying expectation depending on their prior satisfaction. While it is arguably clear that SWS can play a role in improving practices for high-level orchestration and traceability of scientific workflows, we explore in this paper the granularity at which activities can be usefully defined. We describe a case study, the calibration of a model, wrapping the components of a modelling framework (TIME) from the Trident and Kepler SWS. The problem is decomposed in several workflows comprising activities of differing granularities. We assess each approach against a set of criteria such as runtime performance and flexibility, discuss the feasibility and trade-off. The main findings are the design benefits stemming from having to clearly identify separate activities in the process, and that the difficulty of decomposing the optimization problem into finer-grained activities increases most markedly when needing iterative control flow capabilities.