Monday, September 6, 2010

WCC tuning

Most of the optimization for WCC can be accomplished by taking the previously mentioned
steps of planning transactions carefully. However, we make the following WCC-specific
configuration recommendations:
Disable suspect duplicate processing for exceptional workload cases that might not
benefit from it. This is especially recommended for initial WCC data migration or loading
because not all of the customer data is present for the suspect duplicate processing to
make an accurate suspect list during most of the data loading process. Instead, use the
evergreening process after the data is loaded. Consult the WebSphere Customer Center
Understanding and Planning Guide for more information about evergreening data.
Disable TAIL if you do not require that level of auditing in your implementation. In most
cases, also disable TAIL for initial WCC data migration or loading.
Use the smart inquiries feature whenever possible.
Disable WCC performance tracking for regular production workloads. Instead consider
using the services activity monitoring feature of WCC (available in WCC version 7.0 or
later).
Reduce WCC logging levels to the minimum required for your infrastructure.
Use the search exclusion feature (available in WCC version 7.0 or later) where appropriate
to avoid producing excessively large results.

No comments:

Post a Comment