Hi all,
I have studied on the Change Data Capture in Katmai. That is new with Katmai. I wonder how it is implemented actually. Does it use a process like service broker? I hope it does not affect the performance of the ongoing processes or transactions on the database engine.
I worked on a sample on my Katmai installation and took my notes at http://www.kodyaz.com/articles/change-data-capture.aspx
I find the tables and functions a little bit confusing. But after working on a sample it makes it easier to understand the structure.
With the preliminary information I have now this is retrieved from the transaction log.
The 'inside' information is not yet available I'm afraid :-) There is a nice schema in the BOL though.
Everything has a cost. The question is: "is the performance impact acceptable on my environment" and this can only be determined by tests on your specific environment.
WesleyB
Visit my SQL Server weblog @. http://dis4ea.blogspot.com
|||The CDC will definitely add some extra disk activity to the system, and thus I wonder if there will be an option to specify a filegroup to place the CDC data on, to improve performance of course.|||Yes, you have control at the physical level. The only thing I don't like at all is the way to query it. It exposes the guts of the system at the logical level. This is all fine for debugging (as a DBA, access to the LSN is a goodie) but it should not be exposed to query/develop againts it. Mixing the logical and the physical is bad design.
