Results 1 to 3 of 3
  1. #1
    Join Date
    Dec 2009

    Unanswered: SQL 2005 - server and client question

    We are wanting to move to one central sql database and have our 50 locations connect to the server at a central place, rather then having sql loaded at each location and replicate the data everywhere.

    We tested this out today. It connects just fine but it SLLLOOOWWW to do everything. Is there any tricks to speed this up? All internet is pretty quick connections.


  2. #2
    Join Date
    Aug 2009

    one thing you can try is to speak with the network device admins to modify the switch configurations. It is possible to give the data packages of this server more priority to others (especially Internet), this can fasten the access a lot.
    Another thing is of course to check the client application and optimize it. It's very often that forms uses complete and very long tables to be filtered locally or other such things which can slow down the performance in every scenario.



  3. #3
    Join Date
    May 2009
    Provided Answers: 1
    halen99, The slowdown could also be the way the applications were written. If every user had a local database this wouldn't matter too much. But if the applications do all the 'heavy' SQL work, you could be transferring a lot a lot of row that are being filtered out. Instead, the programs should be calling Stored Procedures on the Server to do the bulk of the SQL work. For example, if a program:

    Start program.
    Connect to Server.
    Select cols.. From Tab1 inner join Tab2 on join-predicate Where filter-list
    Process results
    End program.

    It would have to retrieve all the rows from the tables, transmit them over the network, filter them in the program locally and retain just the rows that the SQL needs. This could be a LOT of network traffic. Instead it should be more like:

    Start Program.
    Connect to Server.
    Call Stored Procedure on Server.
    Process result set.
    End Program.

    This way a request is sent to the Server, the Filter of the SQL is completed at the Server and only the (hopefully) small number of rows will be transmitted over the network back to the program.

    This (especially in conjunction with Bitsqueezer suggestion) could have a very big impact on performance.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts