Re: S&P performance DB/2 vs. SQL Server as TDWH

hi,

might be down to the fact that partitioning has been implemented with the new TDW, check the warehouse plus associated agents configuration 

http://pic.dhe.ibm.com/infocenter/tivihelp/v61r1/topic/com.ibm.itm.doc_6.3/install/tdw_rangepart.htm
https://www.ibm.com/developerworks/community/blogs/0587adbc-8477-431f-8c68-9226adea11ed/entry/how_to_check_the_table_range_partitioning_in_a_tdw_database?lang=en


 
Regards 
Justin Burland

Mobile - +44 7906 562118
email - [email protected]


________________________________
 From: "[email protected]" <[email protected]>
To: [email protected] 
Sent: Monday, 19 August 2013, 14:01
Subject: RE: [TME10] S&P performance DB/2 vs. SQL Server as TDWH
 


 
Hi
It is the version that came with itm6, 9.7.400.
Yes it’s propably because of the fact that I have set up a maint. Schema with reorgs etc. Don’t recall that ever happened with SQL.
 
Still, it’s such a huge difference. I was almost chocked at the fact. And it has been running for 3-4 months now.
 
/Gus
 
 
From:[email protected] [mailto:[email protected]] On Behalf Of [email protected]
Sent: den 19 augusti 2013 14:54
To: Discussion list for Tivoli product and Tivoli Ready products.
Subject: Re: [TME10] S&P performance DB/2 vs. SQL Server as TDWH
 
What version of DB2 did you go to, the new 10.5 BLU? It has been massively tuned for OLAP workloads and would explain it.  In general, DB2 outperforms Oracle and SQL Server with specific tuning, funny that not many realize this.  With pinpoint tuning for both OLAP and OLTP trnasactions the gap widens. And for appliances (i.e., Netezza, OLAP only)...Fagetaboutit. ;)
But more specifically, what likely contributed to such a magnitude runtime improvement is that the indexes were rebuilt during the data load into DB2. The tables also were created with minimum fragmentation and the optimizer stats accurate.  It will be interesting to see how it's running in a couple of months.
-Toben 
On Aug 19, 2013 7:08 AM, <[email protected]> wrote:
Hi.

We recently changed DB engine from SQL server to DB/2 for our Windows TDWH.
S&P has always been run on the same box which is hosting the DB.

On the same W2K8 12 GB ram Quad 2.28Ghz box, we went from 9 hours S&P to 9 minutes!!
How can this be possible? Has IBM tuned the S&P agent extremely well for DB/2.

Initally I thought that times would increase along with the DB content, but now we have been running as long as our longest summarization and we still see around 10 min.s

/Gus



_______________________________________________
TME10 mailing list
[email protected]
Unsubscribe:[email protected]
_______________________________________________
TME10 mailing list
[email protected]
Unsubscribe:[email protected]

_______________________________________________
TME10 mailing list
[email protected]
Unsubscribe:[email protected]