Mediation frameworks present a cost-effective solution to integrating distributed heterogeneous data sources. Mediators are middleware in which a new layer is introduced between applications and data sources, so changes in either application layer or data sources layer will not affect the other. The mediation layer adds value to systems such as data conversion, integration, and presentation. Before data can be integrated, it must be found: a simple solution to this is to maintain a repository for the schemas of the data sources. However, this approach affects the reliability and scalability of the system. This research is to design a large scale, reliable mediation architecture based on a skip-list distributed hash table algorithm.