From Bill Porter, Marketing Director, White Cross Systems, Bracknell, Berkshire RG12 1RB.
I read with wry amusement Abigail Waraker’s account of the CAPPS conference (CI No 2,531). It seemed to me to be hugely out of context. I can only assume that the conference speakers were not of this world, or perhaps this time. The perspective is that of a 1960s Cobol programmer who has yet to catch up with program-data independence, database, relational theory and SQL, never mind client-server concepts. Yes, the hard part is the software. To make it work reliably and with the full benifit of parallelism, it has to be designed and written from scratch. What’s worse, users are not interested in knowing whether their computer is parallel or not. Could this be a clue as to why the opportunity for the commercial parallel systems must be in the database? You bet. SQL is a non-procedural language. It specifies what the user wants, not how to do it. Lots of people use SQL, many use it without knowing it. Tools and applications generate it and then pass it through a standard interface like the X/Open CLI and Open Data Base Connectivity. So here’s the great opportunity for parallel database systems. We can write incredibly powerful new software on parallel hardware that can do wonderful knowledge enablement things for business people, without them hardly realising it. So commercial massively parallel processing does database, big time, if the software problem can be cracked – as Teradata did in its way 10+ years ago, and as White Cross Systems has done again (but better). Problem solved. But there are many people who say, Database – is that all?. Sigh. Don’t I hear that the relational database market for software alone is worth some $5 billion? And that the relational software market is growing at some 20% per annum? Isn’t it true that 70% of all servers are for database? Oh, by the way, in case you hadn’t heard, there’s a massive interest in Data Warehousing. Reputable market researchers predict a large market for commercial massively parallel processing, growing at some 30% or more a year. By talking to the users that can or do benifit from it, its not hard to understand why. Maybe Abigail Waraker should listen to them, instead of the academics’ algorithm hang-ups.
