Top 5 relational data warehouse performance improvements you would like to see in next version of SQL Server

  • Article

Hello Everyone,

We are currently looking at top performance issues for various workloads and how we can improve those in next version of SQL Server. We are collecting feedback from various customer sources. I would like to extend an invitation to readers here for feedback on top 5 relational data warehouse performance improvements you would like to see in SQL Server.

Please bear the following in mind:

1. Briefly describe the relational data warehouse workload or scenario or application. More details you can provide for us the better.

2. Prioritize the list of performance improvements in order of importance or impact to your workload. You can also just highlight the top 5 performance problems with running your relational data warehouse workload in SQL Server

3. Feel free to suggest any feature(s) that will help in improving the performance of the particular data warehouse workload or scenario

I will consolidate the feedback and post summary version of it as a link in this post.

Thanks
Umachandar

Comments

  • Anonymous
    September 21, 2006
    The comment has been removed

  • Anonymous
    September 25, 2006
    Please solve SQL 2000 problem 469591 ("When using sp_cursoropen with an Order By Clause we Generate an Inefficient plan ")

  • Anonymous
    September 26, 2006
    please ignore my posting from September 25, 2006 11:43 AM
    wrong category

  • Anonymous
    September 26, 2006
    If you use SQL Server and have a great suggestion for performance improvements to the engine then post...

  • Anonymous
    September 27, 2006
    The comment has been removed

  • Anonymous
    September 28, 2006
    Can we have some unsigned int datatypes?

  • Anonymous
    September 28, 2006
    Can we have some unsigned int datatypes?

  • Anonymous
    September 28, 2006
    Not sure if I can call this as DW improvements.  But for sure I can say, if you guys can provide a URL where you guys can post "Case Studies" explicitly describing

    1. Nature of Problem
    2. What was accomplihsed
    3. Technology used
    4. What was the performance growth
    5. User base (# of users)
    6. Network Topology
    7.....

    May be this information is missing across web & I have been searching a lot on this, but noluck

  • Anonymous
    October 12, 2006
    a) Can we have some unsigned int datatypes? b) Can we have some unsigned tinyint datatypes? c) Can we have some unsigned smallint datatypes? d) seperate date and time datatypes e) ability to remove custom datatypes from a large existing database easily

  • Anonymous
    October 12, 2006
    The comment has been removed

  • Anonymous
    October 16, 2006
    General note. Please do not post general feature requests. We are looking specifically for performance issues only. For example, there were several posts asking about unsigned int data types. Why do you need unsigned data types? How do you think it will improve performance of your application? Also, we are asking for relational data warehouse performance improvements only not Analysis Services. Thanks Umachandar

  • Anonymous
    October 26, 2006
    The comment has been removed

  • Anonymous
    December 19, 2006
    There's been some interesting questions by users on the SQL Server Engine Tips blog. While almost all of them seem to be off topic, there are some interesting ideas. One of the fun ones was from Dave asking "Can we have some unsigned int data types?" So assuming Dave is wanting unsigned integers for this reason, I created a blog entry on how you would create unsigned integer user datatypes in SQL Server 2005...