3 Replies Latest reply on Jun 4, 2013 1:13 PM by Victor Armbrust

    X3 overhead due to WriteBack


      Does anyone know whether there is any overhead in X3 when using the new WriteBack policy and use FlashCache for write operations?. We are planning to migrate from X2 model to X3 and ACS told me that there is a little overhead but I don´t know what is the reason and the percentage for this overhead.

        • 1. Re: X3 overhead due to WriteBack
          Hi Friend

          There are four methods are available in Write-Back Flash Cache.

          1. Write-through Read Path
          2. Write-back Read Path
          3. Write-through Write Path
          4. Write-back Write Path

          a. Flash cache provides the enhanced read and write performance.
          b. This means writes first happen on Flash cache and the database gives acknowledgment.
          c. Cell software then writes them back onto hard disk when it finds it necessary.

          It will improve performance for Applications like most "Insert" operations. Before configuration we have to configure ASM disk, Flash Cache carefully to avoid performance bottle necks

          Hope it helps..

          • 2. Re: X3 overhead due to WriteBack

            thank you for your answer, very interesting, but, do you know what´s is the reason for the overhead ?, that´s the true problem.

            • 3. Re: X3 overhead due to WriteBack
              Victor Armbrust
              LaserSoft said all that you need to know (lol). Perfect

              I had implemented WriteBack in several customers at ACS, and I really don't see any kind of bottleneck till now. all implementation works fine (for X2-2 and X3-2). Last week I had implemented a X3-2 Quarter Rack with Writeback with no problem, all FlashCache operations work fine (Off course, X3-2 has more FlashCache memory than X2-2, but I really don't see any kind of bottlenecks)


              Edited by: Victor Armbrust on 04/06/2013 10:13