MEDWRAP: Consistent view maintenance over distributed multi-relation sources. Warehouses today extract information from several sources, each with multiple relations. Incremental View Maintenance (VM) of warehouses in such environments faces the problem of concurrency conflicts due to simultaneous relational updates occurring within and across these (semi-autonomous) sources. Existing VM algorithms only partially solve this issue. Some like ECA and CCA assume a single-source warehouse, while others like Strobe and SWEEP assume a multi-source environment with only one relation per source. However, in practice data sources have multiple relations in one schema. In this paper, we propose a solution called MEDWRAP that applies two-layered compensation. It resolves concurrency conflicts by using single-source compensation at each source wrapper and multi-source compensation at the mediator. We show that this achieves correct and consistent view maintenance. Not requiring intermediate views to be stored at the wrapper, MEDWRAP is space-efficient, a highly desirable feature, given the ever increasing size of modern warehouses.