What are the various parameters needed for ETL performance

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
vikas gupta
Participant
Posts: 30
Joined: Wed Mar 24, 2004 1:28 am

What are the various parameters needed for ETL performance

Post by vikas gupta »

What are the various parameters needed for ETL performance
ashwin141
Participant
Posts: 95
Joined: Wed Aug 24, 2005 2:26 am
Location: London, UK

Re: What are the various parameters needed for ETL performan

Post by ashwin141 »

vikas gupta wrote:What are the various parameters needed for ETL performance
Hi Vikas

There are many parameters on which your ETL performance depends. I can list some of them here.

The flexibility of the architecture. It shouldn't be tool dependent.
Which database engine will be used for the transformations?
What happens to ETL developments if the physical address of a server changes or if a database name changes?
How you promote code from development to production.
Complexity of business rules.
Number of dimensions and their nature.
Platform.

There is a long list. I hope it helps u.

Regards
Ashwin
kumar_s
Charter Member
Charter Member
Posts: 5245
Joined: Thu Jun 16, 2005 11:00 pm

Post by kumar_s »

Also check for APT_BUFFER_MAXIMUM_MEMORY, APT_BUFFER_FREE_RUN, Buffering policies, Database parameters.
Impossible doesn't mean 'it is not possible' actually means... 'NOBODY HAS DONE IT SO FAR'
DSguru2B
Charter Member
Charter Member
Posts: 6854
Joined: Wed Feb 09, 2005 3:44 pm
Location: Houston, TX

Post by DSguru2B »

There are a lot of posts covered on ETL performance. I just did a search and got several posts. Also, your answer depends upon what do you mean by performance.
Creativity is allowing yourself to make mistakes. Art is knowing which ones to keep.
ArndW
Participant
Posts: 16318
Joined: Tue Nov 16, 2004 9:08 am
Location: Germany
Contact:

Post by ArndW »

The original question is like asking "What can I do to tune my vehicle" in a car forum and can't really be answered, at least directly.

Reading and writing data takes a long time. Try to read and write and "land" data to disk as few times as possible. This approach (while not always correct) usually works pretty well for a starting point.
ray.wurlod
Participant
Posts: 54607
Joined: Wed Oct 23, 2002 10:52 pm
Location: Sydney, Australia
Contact:

Re: What are the various parameters needed for ETL performan

Post by ray.wurlod »

vikas gupta wrote:What are the various parameters needed for ETL performance
The first thing you need is some definition of "performance". In ETL the key performance indicator is usually the ability to meet some time window in which access to the databases in question is permissible. Nothing else really matters.

For example, you may be able to process the same volume of data faster just by throwing more processing nodes at it.

Intelligent design and some awareness of which techniques are more efficient than others. Some knowledge of how the ETL tool interacts with data sources (and what alternatives are available) is also vital.
IBM Software Services Group
Any contribution to this forum is my own opinion and does not necessarily reflect any position that IBM may hold.
develop
Premium Member
Premium Member
Posts: 7
Joined: Fri May 26, 2006 4:34 pm

...

Post by develop »

..
develop
Premium Member
Premium Member
Posts: 7
Joined: Fri May 26, 2006 4:34 pm

...

Post by develop »

..
Last edited by develop on Thu Jun 01, 2006 2:42 pm, edited 1 time in total.
DSguru2B
Charter Member
Charter Member
Posts: 6854
Joined: Wed Feb 09, 2005 3:44 pm
Location: Houston, TX

Post by DSguru2B »

Arent you having fun with those dots :twisted:
Creativity is allowing yourself to make mistakes. Art is knowing which ones to keep.
Post Reply