StorHouse/Relational Manager (RM) - Active Storage Hierarchy Database System and Applications

This paper describes how database systems can use and exploit a cost-effective active storage hierarchy. By active storage hierarchy we mean a database system that uses all storage media (i.e. optical, tape, and disk) to store and retrieve data and not just disk. We describe and emphasize the active part, whereby all storage types are used to store raw data that is converted to strategic business information. We describe an evolution to the Data Warehouse concept, called Atomic Data Store, whereby atomic data is stored in the database system. Atomic data is defined as storing all the historic data values and executing queries against the historic queries. We also describe a Data Warehouse information collection, flow and central data store Hub-and-Spoke architecture, used to feed data into Data Marts. We also describe a commercial product; StorHouse/Relational Manager (RM). RM is a commercial relational database system that executes SQL queries directly against data stored on the storage hierarchy (i.e. tape, optical, disk). We conclude with a brief overview of a real world AT&T Call Detail Warehouse (CDW) case study. 1.0 Introduction Commercial Database Management Systems (DBMS) have evolved and been developed to diverse and ubiquitous range of applications. DBMS have been based on hierarchical, network, relational, objectoriented and the new emerging object/relational database model. With few exceptions these database systems and applications primarily use disk media as their storage. Hierarchical Storage Management (HSM) is used by some of these applications to exploit some of the benefits of cost-effective optical storage systems. We propose and analyze that database systems use and exploit a complete active storage hierarchy (i.e. tape, optical, and disk). The key proposal is that active data be also stored, queried and analyzed on tape farm libraries and optical jukeboxes. Figure 1 shows the cost, performance, size and reliability considerations. In this paper, we use the term active storage hierarchy when (say SQL) queries execute against data stored on diverse media.