Results 1 to 3 of 3

Thread: How to do the performance tuning at various levels

  1. #1
    Junior Member
    Join Date
    Oct 2006
    Answers
    1

    How to do the performance tuning at various levels

    Can any one tell me how to do the performance tuning at Source level, Target level, Mapping level and Session

    Thanks
    Dev


  2. #2
    Junior Member
    Join Date
    Apr 2007
    Answers
    4

    Re: How to do the performance tuning at various levels

    You can improve performance at source leve using sqlquery and sorted input
    and another way is to keep a filter transformation after sql transformation
    Using partition type key range at source level.
    Target level writing to file loads data faster than database. and use key range partition type at target level.
    Tranasformations permformance can be improved using look up with dynamic cache and keeping erlier filter using Aggregator with sorted input ,remove unwanted ports from lookup,filter and other trasnformations
    Session level Partioning helps u to improve performance.reading the session log details give where the bottleneck occured by cheking busy process and idle process. if the busy percntage is more than idle percentage it is a bottleneck.
    I hope these are bottlenecks occured in informatica and using above technics improve performance


  3. #3
    Junior Member
    Join Date
    Feb 2007
    Answers
    4

    Re: How to do the performance tuning at various levels


    Hi Praveen,
    First we have to identify the bottlenecks and then optimize.

    1. Target : Load the data to a table and a flat file. If there is a time difference we can say that there is a target bottleneck.

    Use target load type BULK.
    Drop indexes and key constraints before loading and re-built them after the loading.

    2. Source: For flat file sources mostly bottlenecks don't araise. If so increase the no. of bytes power center reads per line (default 1024 bytes) in the session properties.
    If source is a table, we can identify using filter conditions and sql qery execution. Use filter transformation and set the condition as FALSE and observe the time.
    Execute the sql query in the source database, observe the time taken for retrieving the rows.
    Optimize the sql query in the SQL override and use conditional filters.
    3.Mapping : We can identify the bottlenecks using conditional filters.
    More lookups in the mapping reduces performance.
    Avoid unnecessary data type conversions.
    Remove unnecessary pipe lines and output ports.
    4.Session: Use "Collect performance details" option in the session properties, so that we can get performance details log.In that any value other than zero for read to disk and write to disk, then there is a bottlneck.
    Increase the commit intervals and cache size.
    Set the tracing level Normal or Terse.
    5.System : Using some tools like Task manager we can identify.
    Improve the system configuration and network.

    If any corrections needed give reply.


    Thanks & Regards,
    G Gopi Krishna.




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
About us
Applying for a job can be a stressful and frustrating experience, especially for someone who has never done it before. Considering that you are competing for the position with a at least a dozen other applicants, it is imperative that you thoroughly prepare for the job interview, in order to stand a good chance of getting hired. That's where GeekInterview can help.
Interact