White Papers
Deploying Microsoft SQL Server 2008 based Decision Support Systems using Dell EqualLogic 10GbE iSCSI Storage 2
2 Understanding DSS Storage Requirements
DSS applications are typically designed to support complex analytical query activities using very large
data sets. The queries executed on a DSS database typically take a long time to complete and usually
require processing large amounts of data. A DSS query may fetch millions of records from the
database for processing. To support these queries the server reads table data from the storage devices.
The data I/O pattern from storage to server is predominantly sequential and usually consists of large
blocks – typically 512KB or 1MB in size.
It is an important best practice to run DSS workloads on a physically separate system that is optimized
for DSS query I/O activity. This will prevent DSS workloads from interfering with the operation of other
workloads.
2.1 Strategy for Supporting DSS I/O Requirements
The large sequential I/O patterns in DSS queries require careful system design to insure that the
servers, interconnect fabric and storage arrays are able to meet performance and capacity
requirements. The overall performance of the DSS solution is determined by the performance
characteristics of each component in the system: from the SQL based DSS application driving the I/O
workload down to the individual disk drives in the storage area network (SAN) that store the data. The
best practices presented in this paper focus on all of the component layers presented in Figure 1.
Figure1:ComponentsinaDSSSolutionstack