February 4, 2009

Draft slides on how to select an analytic DBMS

I need to finalize an already-too-long slide deck on how to select an analytic DBMS by late Thursday night.  Anybody see something I’m overlooking, or just plain got wrong?

Edit: The slides have now been finalized.

Comments

12 Responses to “Draft slides on how to select an analytic DBMS”

  1. Dave Menninger on February 4th, 2009 6:51 pm

    I think you should also distinguish column-oriented vendors by their MPP capabilities or lack thereof. My very quick review of the deck (30 sec) it seems that you only associate MPP with row-based vendors.

  2. Curt Monash on February 4th, 2009 6:57 pm

    Dave,

    I’d rather not be an extremist about that. On the row-based side, MPP is pretty well correlated with size of database handled well (SAS excepted, but they’re a super-special case). On the columnar side, that’s not so true, due to compression and due to the fact that Exasol is so RAM-centric. (We could add ParAccel’s TPC-Hs to the litany.)

    It definitely bears saying more about, but I don’t think I’ll actually split the categories.

    Good catch even so.

    Thanks!

    CAM

  3. AV on February 5th, 2009 1:37 am

    Great post.

    You mention manageability and availability in the same item. Is there a manageability checklist which is distinct from availability?

    Should the buyer focus on day-to-day stability of long standing reports? How does the DBA work with slowdown of queries? Does he run explain plan? How many stats collection jobs does he have to run nightly? These are unglamorous manageability activities but are very, very time consuming.

    Of course, in an ideal world, the queries always run faster and faster. And the world wouldn’t need DBAs.

  4. Imre on February 5th, 2009 6:40 am

    Nice Post!

    +1 factor: the “ETL Tool” …
    +1 keywords: “Open Source” vs . “Closed Source”

  5. D on February 5th, 2009 10:25 am

    Great presentation.

    A couple of points-

    I think it is important to understand functionality, cost and performance requirements and determine the intersection of these. For example, I have encountered folks who just need a large amount of historical data available for occasional analysis with very few concurrent users (if at all). If availability and cost (and not performance) is a concern, they can get by with a cheaper solution. (Of course, once it is live, chances are it will receive more use than originally expected over time.) The flip side of this is where money is no object and performance is crucial, they will probably look at a different set of vendors.

    Another minor comment is that POCs do take time and effort to manage with the vendors, so one may want to first winnow down the list after all instead of doing many.

  6. John Sichi on February 5th, 2009 8:22 pm

    Hi Curt,

    It would be nice to see LucidDB in there eventually. As an open source project with no marketing department behind it, I guess we’re under your radar, but I’d be glad to fill you in when you have time.

    JVS

  7. Satil Hecht on February 5th, 2009 11:18 pm

    Hi Curt, Nice presentation. Couple of comments you may want to consider:
    - There are differences from one vendor to the other when it comes to the duration and cost to get up and running. Appliances like the IBM BCU and Teradata help remove the risk in integrating the right components, and ensuring that you have the right balance and configuration for success when you go live.
    - Another consideration is integration with other components of the data warehouse like ETL, archiving solutions, and BI tools. We found some vendors either lacked the support needed, or left majority of the integration for us to figure out.

    Thanks!

  8. 分析型数据仓库选型 | Alex的个人Blog on February 6th, 2009 1:08 am
  9. Curt Monash on February 6th, 2009 2:00 am

    Satil,

    You’re of course right. But I don’t want to talk about the compatibility issues at any length, because they really are a matter of checklists (backed up by reference checks if need be).

    Maybe I need to add a slide at the end that’s a “Checklist of checklists”.

  10. Final (for now) slides on how to select a data warehouse DBMS | DBMS2 -- DataBase Management System Services on February 6th, 2009 9:15 am

    [...] now posted a final version of the slide deck* I first posted Wednesday.  And I do mean final; TDWI likes its slide decks locked down weeks in advance, because they go to [...]

  11. My TDWI Night School course Wednesday night | DBMS2 -- DataBase Management System Services on February 18th, 2009 3:59 pm

    [...] be holding forth Wednesday night on How to Select an Analytic DBMS.  I’ve already posted the slides. Share: These icons link to social bookmarking sites where readers can share and discover new web [...]

  12. Even more final version of my TDWI slide deck | DBMS2 -- DataBase Management System Services on February 25th, 2009 9:09 pm

    [...] My TDWI talk on How to Select an Analytic DBMS starts in less than an hour.  So the latest version of my slide deck should prove truly final, unlike my prior two. [...]

Leave a Reply




Feed: DBMS (database management system), DW (data warehousing), BI (business intelligence), and analytics technology Subscribe to the Monash Research feed via RSS or email:

Login

Search our blogs and white papers

Monash Research blogs

User consulting

Building a short list? Refining your strategic plan? We can help.

Vendor advisory

We tell vendors what's happening -- and, more important, what they should do about it.

Monash Research highlights

Learn about white papers, webcasts, and blog highlights, by RSS or email.