Tech Tweets

    follow me on Twitter
    AddThis Social Bookmark Button
    Blog powered by Typepad

    March 2021

    Sun Mon Tue Wed Thu Fri Sat
      1 2 3 4 5 6
    7 8 9 10 11 12 13
    14 15 16 17 18 19 20
    21 22 23 24 25 26 27
    28 29 30 31      

    Become a Fan

    « Wokcano Meetup | Main | Eigenbase: First Look »

    November 16, 2010

    Comments

    victor

    Hi,

    When you speak about Cloud, OLAP and big data (interesting subjects) , with this week Amazon Cloud announce, I think of a Palo OLAP cube into a Amazon GPU instance.
    Do you know this?
    http://news.softpedia.com/news/NVIDIA-GPU-Supercomputing-Available-via-Amazon-Cloud-Offering-166521.shtml
    http://www.jedox.com/en/products/palo-gpu-accelerator.html


    Victor

    Cobb

    I heard about the GPU announcement this week. It immediately made me think about SeaMicro because they too are using a different chipset to enable higher performance computing. 


    It looks like Palo is hooked up on many levels, and I like the idea of a memory resident OLAP. Now that Im thinking a little bit about ROLAP and various necessities for creating aggregate tables, Im wondering how and why some of that remains necessary. 

    Let me go off on a little tangent. For about six or seven years, I kept track of the amount of data I could get an Essbase cube to aggregate per hour. I stopped counting maybe five years ago when it got around 10GB. And Ive also kept track of the number of dimensions that are generally human cognizable. So when I think about the amount of data presented to a small set of users, it makes no sense to me that there is anything that should require a great deal of time to compute. It would take me a very long time to visualize what amounts to 10GB of data - certainly more than an hour, so if Im just trying to see a particularly small set why shouldnt my response time be anything less than instantaneous. 

    Now theoretically, the ability to recognize all of the ancestors and dependents measures for any known visualization set, ie what a reasonable user would query in a known standard report - should be deterministic. It should be a simple matter to specify every bit of info I need. So the first thing I would think of, especially in a planning application with writeback that a user wants immediately updated, is a machine generated profile of sorts. It seems that this is something that Essbases ASO should do well. You could certainly reverse-engineer something like that. The lower in the hierarchy the set is, the easier to define the ancestor set. But any ROLAP should have the same properties. At some level of the API, MDX could be generated to cache that set of aggregations. I should be able, basically, to materialize any set of sub-cube data I want, and then having done so put them on a fast track for update - given that I have a clean/dirty facility - by knowing the standardized reports. 

    Verify your Comment

    Previewing your Comment

    This is only a preview. Your comment has not yet been posted.

    Working...
    Your comment could not be posted. Error type:
    Your comment has been posted. Post another comment

    The letters and numbers you entered did not match the image. Please try again.

    As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

    Having trouble reading this image? View an alternate.

    Working...

    Post a comment

    Your Information

    (Name is required. Email address will not be displayed with the comment.)