Online analysis of large volumes of data consists an important part of today's business and scientific applications. Storing structured data in data warehouses following a multidimensional storage model provides efficient and reliable access to it. Distributed systems have become the first choice to cope with the increasing volume and complexity of data warehouses. In order keep up with this trend, large quantities of interconnected computing and storage resources are now being integrated into computing grids. Adapting data warehouse concepts to grids however requires decentralized management of their traditionally centralized storage and query execution model. The grid services architecture we present in this paper therefore uses a unique data identification to implement grid-wide data localization based on local indexes in combination with a catalog service. These structures integrate information on both materialized and computable data for cost‑optimized distributed query execution.