1230
Part IX: Business Intelligence
Although the General node properties are the main focus of the Analysis Server Properties,
the Language/Collation and Security settings may also be changed using this
properties dialog. Users or groups entered in the Security section will become Analysis
Services Server administrators.Advanced SSAS Deployments
Analysis Services supports quite a number of options to assist in meeting processing or
querying goals. Key to the architecture of Analysis Services is that it natively supports
server failover clustering.Processing can be done on one instance of Analysis Services, and the processed data may
then be copied to other instances used for querying. Many different ways to copy the data
exist including but not limited to SAN Snapshots, Storage Mirrors, XCopy, and Analysis
Services Cube Synchronization.Querying can be managed through the use of several identical Analysis Services instances
that are clustered and placed behind a Network Load Balancing (NLB). More details are avail-
able in the SQLCAT article Analysis Services Load Balancing Solution located at http://
sqlcat.com/sqlcat/b/technicalnotes/archive/2010/02/08/aslb-setup.aspx.Reviewing Query Performance with SQL Profi ler
Effi cient and fast queries are the cornerstone to a successful Analysis Services solution. Key
for delivering the expected performance is ensuring that the system is optimized.The main tool used to monitor the performance of Analysis Services is SQL Server Profi ler.
SQL Server Profi ler contains a vast array of Analysis Services events that can be used to
monitor performance, determine poorly performing MDX statements, step through MDX
statements, replay events on another system, and audit activity. This section provides a
brief introduction to using SQL Server Profi ler for monitoring Analysis Services.Event categories that can prove useful for tracing the Analysis Services server activity
and investigating the performance of MDX queries submitted to the server to process user
requests include:■ Command events: These provide insight into the actual types of statements issued
to perform actions.
■ Discovery events: These detail requests for metadata about server objects such as
open connections.
■ Error and Warning events: These events alert the developer to issues with build-
ing and deploying Analysis Services databases.
■ Notification events: These events notify the user through tracing that certain
query actions are taking place.c54.indd 1230c54.indd 1230 7/31/2012 10:31:05 AM7/31/2012 10:31:05 AM
http://www.it-ebooks.info