07.02.2013 Views

Best Practices for SAP BI using DB2 9 for z/OS - IBM Redbooks

Best Practices for SAP BI using DB2 9 for z/OS - IBM Redbooks

Best Practices for SAP BI using DB2 9 for z/OS - IBM Redbooks

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

► Load impact<br />

– Loading data into the InfoCube will be faster, because it usually involves<br />

inserting into an empty or small F- fact table and its dimension tables.<br />

– Additionally, if indexes on the F- fact table are dropped to improve insert<br />

per<strong>for</strong>mance, then REBUILD INDEX time will be much shorter due to a<br />

smaller F- fact table.<br />

► Administration impact<br />

Important: Lack of RUNSTATS on E and F is the major cause of poor<br />

compression per<strong>for</strong>mance.<br />

– <strong>SAP</strong> compression results in a smaller F- fact table.<br />

– <strong>SAP</strong> compression supports partitioning of the E- fact table.<br />

– Administration of compressed InfoCubes is quicker because a smaller F-<br />

fact table means reduced RUNSTATS runtimes.<br />

– Backups and restores are shorter in situations where <strong>SAP</strong> compression<br />

has resulted in reduced data storage requirements.<br />

– Backups and restores can be shorter where partitioning of the E- fact table<br />

has been per<strong>for</strong>med and the DBA chooses to back up/restore only<br />

changed partitions (since older partitions will not have changed since last<br />

backup).<br />

– Deleting records from the F- fact table is faster due to the smaller size of<br />

the table.<br />

– Updating of aggregates (rollup) is faster when the source InfoCube has<br />

been compressed due to the smaller size of the F- fact table.<br />

– Reorganization is easier on a smaller F- fact table or current partitions only<br />

of the E- fact table.<br />

► Query impact<br />

– Queries against the InfoCube are faster in situations in which<br />

summarization would normally be per<strong>for</strong>med by the application during<br />

query time.<br />

– The summarized data will have fewer records to access.<br />

– Queries will also be accelerated due to E- fact table partitioning where<br />

queries can be defined against partitioning keys, or the query executes in<br />

parallel.<br />

– Query is also enhanced since compression results in less data being<br />

returned.<br />

Chapter 6. <strong>Best</strong> practices <strong>for</strong> InfoCubes 99

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!