2 Replies Latest reply: Sep 17, 2012 3:31 AM by Penky RSS

    Siebel & Oracle Times Ten

    Penky
      Hi All,

      I just want to know if anybody has experience with Siebel and In-Memory Databases like Times Ten. Does it work? Is there huge benefit? Does it make sense at all?
      Our Siebel Database currently is around 120GB in total (DATA + INDEX TS). Since "db file sequential read" is the major wait event and RAM is not very expensive, it might be an option to boost performance.

      Didn't yet check the licensing information, just want to get an opinion from the technical perspective.

      Thanks
      Benny
        • 1. Re: Siebel & Oracle Times Ten
          jiyong
          Does Siebel come with support for Times Ten?
          Or is Times Ten able to "disguise" itself as a traditional Oracle DB?

          I never worked with it, but I see you can use it as an additional cache for a traditional Oracle DB.
          http://www.oracle.com/us/products/database/options/in-memory-database-cache/overview/index.html

          About three years ago we bought a SUN storage solution for our MS SQL DB.
          I think because of some driver issues (iSCSI?), we got SSD's as additional cache on the storage level.
          Also helped a lot for us.

          I haven't seen prices for Times Ten, but I think some good optimizing and tuning with indexes is cheaper.
          Unless you really need sub second transactions, or when you are "abusing" the Siebel DB as an OLAP.
          Well, just noticed the Prices link on the page above.
          $23000 per CPU or $460 per named user (minimum of 25 users).
          I assume more named users will lower the price per user.
          • 2. Re: Siebel & Oracle Times Ten
            Penky
            Hey jiyong,

            Thanks for the information. I don't think Siebel comes with explicit support for TimesTen.
            I am also currently running some test with SSD storage. So far, the results are amazing. An SSD raid configuration in my test environment is performing up to 10x faster on IO intensive queries that our current storage used in production. I think this is the way to go.

            Edited by: Penky on Sep 17, 2012 10:30 AM