Latest data warehousing news

Well, there’s a lot to report.

First, there’s yet another new vendor on the block that, once again, is column-based (that makes 10 now). However, Kickfire is different in that a) it works with MySQL, b) it doesn’t support clustering or MPP and c) it offers an appliance based on a “SQL on a chip” approach. It managed to break the TPC-H benchmark for 100Gb a while ago before being overtaken by ParAccel and then EXASOL. What concerns me most about this is the SQL on a chip: certainly it’s a neat idea, but Calpont couldn’t make it work, which leads me to question whether Kickfire can do so, at least in the longer term.

Perhaps more to the point this whole thing is getting silly – I guess we should start thinking about the great data warehousing bubble – sooner or later it’s going to burst. The question is who will survive, who will crash and burn, and who will get taken over and by whom. I’m thinking about opening a book: any takers?

Secondly, Netezza recently announced general availability of its version 4.5, which includes its compression capabilities as well column-level encryption and other extended security features. More about this in a subsequent article.

Thirdly, though not yet specifically about data warehousing, IBM recently announced the latest version of its SolidDB in-memory database that can be used as a front-end cache to either DB2 or Informix. It seems likely (I am speculating here) that this will be developed further in conjunction with DB2 for data warehousing.

And while on the subject of the InfoSphere Warehouse, the emphasis now is on a modular approach with a foundation module and then multiple data, user and failover modules as required. The particularly neat thing about this is that there is backwards compatibility across these modules so that you don’t have to replace older disk modules (say), when you add some new ones using faster drives. There are also add-on modules for Information Server, Cognos (for which a significant amount of future development is planned, some of it including embedded facilities in DB2), MDM and various third party modules. The last of these are, or will be, certified by IBM for partners such as xkoto, Informatica, Business Objects and so on.

Also of interest is what is called the BI Box. There is an InfoSphere Balanced Warehouse pre-integrated (by Sogeti) with BIReady. Avid readers may recall that BIReady is a competitor to Kalido for rapidly building a data warehouse and then managing change within it, so to have this pre-installed is taking the appliance concept one step further, though I guess it is not intrinsically different from the edge appliances available from the likes of Kognitio and Netezza’s developer network partners.

Finally, I have had an update from HP. The company is now claiming between 25 and 30 customers, most of them large scale systems either replacing Teradata or running alongside Teradata. In order to facilitate the former the company has introduced a migration program with tools for migrating Teradata stored procedures, triggers, DDL and so on, to the NeoView environment. It is likely that the company will develop similar transition tools for other environments in due course, with DB2 probably being next on the list.