Data centric applications often use heterogeneous datasets: some very large while others of moderate size, some highly structured (e.g., relations) while others complex structured (e.g., graphs) or little structured (e.g., log data). Facing them is a variety of storage systems but none of which is the best for all, at all times. We present Estocada, an architecture we are currently developing to efficiently handle highly heterogeneous datasets based on a dynamic set of potentially very different data stores. Estocada provides to the ap- plication layer access to each dataset in its native format, while hosting them internally in a set of potentially overlapping fragments, possibly distributed across heterogeneous stores. At the core of Estocada lie powerful view-based rewriting and view selection algorithms to marry correctness with high performance.