Design

  • Hi,

    I am planning for a database that cater for 4 million users, where each user has around 3kB of data. I read through the documentation in Microsoft website and found the performance stated.

    My superior concern of the performance and capability of MS SQL to support such a huge database. I am wondering, is there any real life production example on how big is the capacity of the DB can hold before the performance get affected? What are the max concurrent users can the db support? what the turnaround time if I post a simple query toward the DB? What the best searching mechanism on 4 mil of records?

    Any help is much appreciated.

    Darian

  • If what you are saying is correct, this database cannot be huge. You have 4M users each holding 3KB data which brings to approximately 1.2GB database and add some space for log file and data file growth. The database got to be small.

    Give me a fish, you feed me for a day! Teach me to fish, you feed me for a life time.

  • Hi,

     

    There is a typo in my thread. Each user should hold 3000 KB of data. Sorry for the information. Much appreciate some advices here.

     

  • Normalize your data, maintain good indexes, and your limitations are going to be within your hardware, backups, number of cpu's, and your available memory.

    Beyond those problems, compare Oracle, MySQL,  Informix, MSSQL & what see what you get??

    All will work based on your available disks & total data size. Each will perform seeks & writes based on your operating system speed...

     


    Regards,

    Coach James

  • http://search.microsoft.com/search/results.aspx?st=b&na=88&View=en-us&qu=vldb

    http://www.microsoft.com/sql/techinfo/administration/2000/scalability.asp

    might be a starting point.

    Apart for this I would buy some external knowledge in form of a consultant with proven experience in planning, design and implementation of such a db and the appropriate hardware for it.

    As for James:

    I like your avatar, but why didn't you mention DB2 in your list?

    --
    Frank Kalis
    Microsoft SQL Server MVP
    Webmaster: http://www.insidesql.org/blogs
    My blog: http://www.insidesql.org/blogs/frankkalis/[/url]

  •  

    Much of the performance and scaling issue is the number of Concurrent users on the system. ie how many will use the system at the same time ( 4 million ? ).

    When it comes to designing large scale applications the important issues is how the system is designed. A bad design will not work however good the hardware / technology is.

    My suggestion is also to bring in help from consultants for

    1. Requirements - Defining the requirements physical & logical for the system.

    2. Architecture - Define the actual physical hardware and the software architecture that can handle requirements ( 1. )

    Honors to James for mentioning Normalization in his suggestion ! Good database design comes with normalization. A denormalized database will seldom see any benefits ( performance or managment wise ).

    I really think that this is a Design question, not a SQL Server question.

    my 2 cents.

    /rockmoose


    You must unlearn what You have learnt

  • Hi Folks,

     

    Thanks for the suggestions.

    If this is concerning about design issue, then I think experienced DBA may render help to me.

    Darian.

  • Darian:

    The comments you received, did not come from novice DBA's!

    Unless you are considering hiring a DBA, I would suggest hiring a consultant!


    Regards,

    Coach James

  • Hi James,

    Sorry if my wording upset you. What I meant in my previous message was hiring experienced DBA to sit down and work through the real solution may help me out

    Darian

  • Ahh. it's nice to see normalisation being mentioned.

    What about SAP or Postgresql on Linux as alternatives ?

    cheers

    dbgeezer

  • SAP ???

    AFAIK, SAP is an application that can run on most major RDBMS.

    Do you mean Abadas or MaxDB (formerly SAPdb)?

    --
    Frank Kalis
    Microsoft SQL Server MVP
    Webmaster: http://www.insidesql.org/blogs
    My blog: http://www.insidesql.org/blogs/frankkalis/[/url]

  • Indeed that is what I meant.

    Thanks for correcting me.

    cheers

    dbgeezer

  • No problem

    Let me correct my own typo

    It's not Abadas, but ADABAS.

    --
    Frank Kalis
    Microsoft SQL Server MVP
    Webmaster: http://www.insidesql.org/blogs
    My blog: http://www.insidesql.org/blogs/frankkalis/[/url]

Viewing 13 posts - 1 through 12 (of 12 total)

You must be logged in to reply to this topic. Login to reply