Introduction Microsoft® SQL Server™ is used to support businesses in a number of mission-critical database processing functions, including online transaction processing (OLTP) and decision support system (DSS) workloads. The speed with which query results are returned is directly impacted by several key factors related to the: • Database—including the nature of the queries (DSS or OLTP, for example) run against the database, the size and complexity of the database, and the type of processing being performed (such as joins or aggregations); • Hardware—commonly the speed and number of CPUs as well as the amount of memory; more rarely, the number of disks; • Software—the efficiency of the application algorithms used to execute the queries. In this paper, we focus on performance testing in DSS environments. DSS workloads, unlike OLTP, are highly resource intensive (see Table 1). Characterized by long transactions with complex queries (either ad hoc or programmed, as with online analytical processing), DSS queries touch large amounts of data (often terabytes in size). As a consequence, DSS queries can potentially saturate both system CPUs and disk bandwidth.