3 Replies Latest reply: May 25, 2004 11:41 AM by 3004 RSS

    Essbase cube 'outer bounds'???

    3004
      Hello - Up to this point, as I have understood it, recommendations are to limit cube size to 12 or so dimensions. Is this still a valid approximation?Also, how correlated is the max number of dimensions with how large those dimensions are? Seems to me that with very thin dimensional structure, you would be able to have mroe dimensions to some point.I have recently built a cube that is 12 dimensions, has been optimized fairly well, and is about 5 GIG of total size. Only calc is a defcalc - total load/calc time is roughly 90 minutes. Retrievals are very fast.I am wondering if I can 'push the envelope' and go even larger. If anyone has any more recent info / insight, I would really appreciate it.Thanks - JMK
        • 1. RE:Essbase cube 'outer bounds'???
          3004
          There is not any documented number of dimensions you can have in an Essbase cube, but you are correct that in reality, most cubes are realistically between 5 and 12 dimensions. Beyond that, you get a lot of inter-dimensional irrelevance and long calc times.From a members perspective, the "unofficial" limit would be 1 million members in a dimension, but there are practical limits that make it less than that. Most client applications would not be able to handle dimensionality of that size, and Excel has a limit of 65,000 rows.Essbase is typically fast on retrieval regardless of how big the cubes are. If your calc times are good and things are working, then don't sweat it too much.Regards,Jade============================Jade ColeSenior BI ConsultantClarity Systemsjcole@claritysystems.comwww.claritysystems.com
          • 2. RE:Essbase cube 'outer bounds'???
            3004
            Thanks, Jade!
            • 3. RE:Essbase cube 'outer bounds'???
              3004
              For your information, most of our cubes consist in 7 or 8 dimensions:Instance of 8-dim cube:1) ACCOUNTS (Dense), 2) TIME (periods) (Dense), 3) TRANSLATED (Currencies, ratios...)(Dense or Sparse),4) AUDIT (Input, Alloc) (Dense or Sparse), 5) PHASE (actuals, budget, simulation..) (Sparse),6) ORGANIZATION (legal/business entities) (Sparse), 7) PRODUCT (Sparse),8) CUSTOMER (Sparse).Those dims are our best tradeoff in terms of calc run times, complexity of calc scripts, retrieval speed...