|
Summer 2001
|
“Bear in mind,” says Jon, “that while young Gates saw the possibility of a single-user operating system that actually allowed you to play Pong, Dick’s product was already a virtual memory based multi-user platform with a common, consistent interface to large, shared databases.
“At the advice of his Chief Licensing Officer at the time, Dick chose to not take the meeting. Can you imagine what the data processing world would be like today if he had, and IBM had anointed him rather than Bill Gates?”
So, Dick, are you watching this?
IBM’s acquisition of Informix appears to have hit the RDBMS market as a bolt from the blue (if you’ll forgive me the awful pun). Nobody seemed to see it coming. For those of us watching from the MultiValue section of the grandstand, it has also reawakened speculation as to the future of the U2 databases.
When Informix took over Ardent, they already had an embarrassment of database products, and seemed to be at a loss what to do with another two, other than milk them for the maintenance revenue. Now IBM has taken on the whole stable, some seven sets of database code, and the poor U2 user is again left wondering where and when the axe is going to fall. This is the third time around for Unidata aficionados - hands up who’s got an ulcer?
There are crucial differences this time, however, which have the potential to make this acquisition the most exciting thing to happen to our market in quite a while.
First off, IBM is huge. Informix was already weakened after its struggles for market dominance against Oracle, and after a return to profitability in six successive quarters was once again starting to look lacklustre. First-quarter results for 2001 showed revenue at $217.1m, down nearly $10m on the previous quarter ($226.8m) and down nearly $33m on the previous year ($250.9m), although with no debt. Perhaps it hoped Ardent - and the DataStage product in particular, which was experiencing phenomenal growth - would provide it with an opportunity to turn things around, so the “legacy” databases of Unidata and Universe were a low priority for them. IBM is a different animal altogether. It has the muscle to actually do something with U2 without straining its core business, if it chooses to - there are advantages to being an $85bn-turnover company, after all. Even if it chooses not to, it has the capacity to keep U2 trundling along nicely for as long as it takes to get the most out of this latest acquisition.
Secondly, IBM is hungry. It has openly stated that it perceives its main competitor to be Oracle, and it is eager to increase its share of the database market. Informix has brought them some 120,000 additional users, and a claimed 13% of the Unix database and data warehousing installed base. Oracle has indicated that it doesn’t think IBM will get much out of the deal, and doesn’t see much threat to its claimed 42% market share, but I suspect that Oracle executives are far less blasé behind the closed doors of their own boardroom.
However, you can’t just buy market share; you have to retain it, consolidate, and expand on it, and you do that by offering something to your customers that the other guy doesn’t, which is where Informix’s portfolio comes in.
Informix as a company always had a reputation for technical innovation, even if the market couldn’t always see the benefits of it, and many of their products contain goodies that IBM has declared it wants to incorporate into future releases of DB2: elements of Cloudscape and DataBlade, for example, and key business intelligence, transaction processing, and object-relational technologies. Given that DB2 already turns in impressive performance and is showing good growth in the Solaris and HP-UX sectors, I don’t think Oracle can afford not to take IBM seriously any longer.
You are probably be wondering how this affects U2 users – I’m getting there. A recent joint Informix/IBM Data Management Solutions white paper setting out the future directions of the Informix stable clearly states that IBM intends to continue to enhance, sell and support the U2 databases and tools, honouring Informix’s plans for new releases, and beyond. It also states that there will be no pressure to move to DB2, and that as long as the customers require certain products, they will continue to be available.
If you were of a cynical turn of mind, you might think that this is merely the usual language of the acquisition, and that the end-of-life announcements will start rolling out all too soon. Undoubtedly, some of the Informix stable will have to be retired; there’s just too much code there to retain indefinitely, and there are several products mentioned in the white paper referred to above whose futures include the words “supply” and “support” but not “develop”. However, there are also lots of ideas in that code, and some very bright people are included in the deal. Janet Perna, head of IBM’s data management division, has described the current state of the RDBMS market as a war for talent, so you can be sure that if IBM are serious about building DB2 into a Universal Database, it will not be discarding any potential ammunition out of hand - and that includes U2.
Those who are cynical for a living, i.e. those whose job it is to ponder and pontificate on the workings of the IT industry, don’t see the future for U2 being at all rosy. In fact, they don’t see a future for U2 at all. They fully expect that the U2 databases, arguably two of the more successful Pick licensees, will be wound down and discarded in short order, useful only for their maintenance revenue (come on, does IBM really need another measly five or six million a year?).
Big Blue itself, by contrast, is at great pains to reassure us that our technological investment is secure, in both its public pronouncements and documents like the Data Management Solutions white paper. At the very worst, it means we can put away the razor blades and hemlock for another year. At the very best? Well, how big are you prepared to dream?
What I found myself dreaming was this: what do you get when you bring together a mature, stable database technology; a powerful company that’s hungry for new weapons in the database wars; a pool of talented developers under the leadership of a determined data management division; a globally-recognised brand; and a British standard shedload of cash?
To be honest, only time will tell. But I do want to say this. I’ve been wondering for 13 years why other folk couldn’t see the benefits of the multidimensional database model (or post-relational, if you apply Codd’s definitions). The relational boys have added the concept of nested tables, after finally realising that real-world data is awkward, complicated stuff that doesn’t easily fit into neat columns and rows. Am I alone in dreaming that maybe, just maybe, we’ve been on the right track all along? And that somebody at IBM, twenty-odd years ago, recognised the potential of Dick Pick’s model, and now a bizarre twist of fate has given them a second chance to bring it the attention it deserves?
I always felt that the MV market was fragmented and too inward-looking, the various vendors wasting energy fighting amongst themselves for a bigger slice of an ever-dwindling pie. End result, nobody on the outside had ever heard of us, and just dismissed us as a relic. jBASE is taking advantage of this by hoovering up the disaffecteds and shepherding them onto a platform that looks like MV, smells like MV, but most certainly ain’t MV, with the promise that they can now take their code anywhere. What I have always wanted to see, though, was recognition of the data model for what it is, not someone dressing it up as something else. IBM is the global brand; they can provide that recognition, at a stroke legitimising our entire market.
All the very best features of the multi-valued data model, all those proven, real-world applications and experienced developers... The potential opportunities here are awesome, unlike anything we have been offered before. With the wealth of GUI and web-enabling tools available to us, there is nowhere we cannot go and nothing we cannot do. It may mean our favourite platforms evolving in unexpected directions, but that will also mean our applications becoming available in new markets, mainstream markets, where they will immediately be made acceptable by the magic three letters. Remember, in the words of the old adage, no-one ever got sacked for buying IBM.
If, along the way, Big Blue happens to wipe the smile off Oracle’s smug, shiny face, well, I think Dick would get a hell of a laugh out of that.
MV News
Last Updated: 20 Jul 2001