Microsoft Corp’s first offering in the online analytical processing server marketplace, still known to the outside world only by its codename Plato, enters its third and final beta phase this month, with a worldwide drop targeted at 30,000 copies, according to its product manager in the UK, Richard Hamblen. As such, it will already outstrip the collective installed base of all existing OLAP servers. Adding them all together and scraping the barrel to include the really small vendors, there are probably no more than 10,000 worlwide, estimates OLAP guru Nigel Pendse.
By Rik Turner
Of course, the comparison is not entirely fair, in that customers don’t have to buy beta copies, but then, the latest indication is that they won’t exactly have to take a specific purchasing decision on Plato once it is on general release either, as it ‘will now be a feature in SQL Server 7.0, as part of the product’, as the company itself explained last month. Now, that does not mean it will be a freebie thrown in, or ‘bundled’, with SS7, Microsoft clarifies, not least because that whole issue is a very sensitive one at the moment, given the litigation working its way through the US courts. It does, however, mean that buyers will be acquiring both servers and that they won’t be separate entities as Hamblen put it. So what will be the effect of Redmond’s entry in earnest into the data warehousing space? Well, the jury’s still out (no pun intended!) on that one. While Pendse predicted a spectacular impact – waxing poetic to refer to the asteroid that destroyed the age of the dinosaur, US data warehousing consultant Larry Greenfield is a more skeptical observer of the market, Greenfield said Microsoft doesn’t want to be a data warehousing vendor. Its principal goal, in his analysis, is to replace MVS and Unix [with NT 5.0] as the business-wide operating system, and to this end, with Plato it is shoring up what could be perceived as a weak point in the attack, namely data warehousing, and as such is defensive, not offensive. Another sector maven, US consultant and author Ralph Kimball, begged to differ. The SQL Server/Plato package is not just an NY pusher, otherwise they wouldn’t have put such quality work into them. Microsoft is keeping quiet on the question of Plato pricing, but the consensus is, in Pendse’s words, that it will be so cheap as to be virtually free, which raises the question of how the competition will react. The British consultant outlined four courses of action. They can fight Microsoft, but so far only Oracle’s doing that, is the first alternative he identified, pointing to Oracle’s championing of the OLAP Council, while most other vendors are adopting Microsoft’s OLEDB for OLAP standard. Alternatively, they can move more into the applications space and out of server technology, he went on, enumerating the second option, which is where Ghentia, Pilot and Comshare are going, and possibly Arbor too with its merger with Hyperion. Indeed, some may eventually buy in their server technology from outside, he continued, pointing to Comshare, which is already using a third-party solution. Pendse’s third alternative is to
jump on the Microsoft bandwagon, endorsing their standard and trying to fill niches they don’t want to, i.e. non-mass market ones, and here he cited Applix, with its TM1 OLAP engine that fully supports Microsoft APIs but with added functionality, and Seagate, with its Holos and Crystal tools. Lastly, he added, there are the start-ups with brand new products, introduced to exploit the new market and the demand for new tools, applications and clients. Companies in this space will be ‘betting you can do a better job from scratch than by adapting,’ and in this context he pointed to Boise, Idaho-based Knosys.
Attractive scenario
It could be argued that the collaborationist approach is a particularly attractive one in the present scenario, given that Microsoft’s basic strategy in data warehousing is to give people the infrastructure, leaving our partners to specialize with their business intelligence tools, in Hamblen’s words. Redmond does not have the armies of in-house consultants in data warehousing that, say, Oracle employs to support its products, and its entire gameplan is clearly one of commoditization and standardization, with its partners providing more niche-specific and/or sophisticated inquiry tools. This line of attack, together with a bargain-basement price proposition, leads Kymball to conclude that in one or two years they’ll be up there with Oracle.
