Database related spending is always a significant cost sink for most CIOs. One component of this spending may be related to maintenance, upgrades, physical hardware costs and other routine items that may not be avoidable. However, a large chunk of the spending may be due to inefficiencies in the data storage mechanism, redundancies in the storage of critical data elements, replication of data access logic and proliferation in the specific data storage technology used. Some of these issues may be addressed by a comprehensive data architecture strategy that attempts to tackle these inconsistencies and redundancies. In this paper, we present architecture for shared data services that may be implemented at an enterprise level to tackle issues associated with data scatter, redundancy and replication. We discuss in detail the various components of such architecture and their applicability towards specific problem areas.