Planning Analytics

Planning Analytics

Get AI-infused integrated business planning

 View Only

  • 1.  Serversizing PAW on Linux

    Posted 06/21/24 02:54 AM

    Dear community,

     

    what is your general recommendation for server sizing for Planning Analytics Workspace (running on Linux RHEL), in terms of

    • Memory
    • Diskspace
    • CPU

    What factors does this depend on?

     

    Is there a rule of thumb how much the sizing should increase for each additional user?

     

    Thank you!

     

    Kind regards

    Rainer Frank

     

     



  • 2.  RE: Serversizing PAW on Linux

    Posted 06/26/24 03:18 PM

    Hi Rainer,

    4 CPU cores with 16-32 GB of memory is a starting point for a Planning Analytics Workspace deployments on Windows or Linux.  In general, Workspace is stateless set of services and you will not need more member based on specific use cases (the exception being extracts from the cube viewer).  The largest single server Workspace deployments I've seen are 64 GB of memory with 32 cores.  If your deployment need to support 100's or 1000's of user you should probably start to examine the distributed deployment option using OpenShift.  Distributed deployments have a minimum of 3 node (typically 4-8 cores with 16-32 GB memory each).  In general you should scale with the concurrent user count.



    ------------------------------
    Stuart King
    IBM Planning Analytics Offering Manager
    ------------------------------