Pop quiz: What is push technology? If you answered Whatever it is those people at Marimba Inc or PointCast Inc do, your answer is sociologically representative, but inaccurate, at least as far as Mark Bowles is concerned. For that gentleman is co-founder and now chief technology officer of Tibco Software Inc, a company that has garnered a lot of attention recently as being a major push vendor (CI No 3,131). But how true is it to say that Tibco is a push company at all? About as true as saying it’s a message oriented middleware company, or even an object request broker supplier, or a custom systems integration house for the financial sector… in other words, yes to all of the above, to lesser or greater degrees.

By Gary Flood

For Tibco is basically pretending to be a trendy push company, and pretending a little less hard to be a still fairly fashionable Common object Request Broker Architecture player, when its heart really lies in message-oriented middleware. To understand this peculiar and almost certainly unique positioning, it helps to bear in mind what kind of an outfit Tibco is. Founded in 1985 as a consulting firm, the company was incubated by a holding company called Teknekron Systems Inc, which looked after it until it sold it lock, stock, and software, to London based international information delivery specialists Reuters Holdings Plc in 1994 for $125m. Tibco grew to its $100m odd current turnover through large scale systems integration contracts, especially to Wall St – its first contract, for a digital trading system for a Boston-based mutual fund company in 1986, resulted in the core software underpinnings of its Tib (The Information Bus) product of today. It toddled on perfectly happily in relative obscurity until the middleware market, thanks to the visibility of products like IBM Corp’s MQSeries, started to acquire an individual identity, making it more necessary to have a defined identity in the marketplace. To achieve this the company split out the middleware (Tib) level of its financial application in 1994 and started to try and build brand awareness around that.

Push furor

That it certainly has done, given the number of partnerships it has won, with companies such as Microsoft Corp, Oracle Corp, VeriSign Inc, CyberCash Inc, Diffusion Inc, and DataChannel Inc, but one gets the feeling Tibco didn’t exactly spring out of bed with excitement at the prospect. We don’t need P.R. in order to get business, Bowles told us at Object World West last week, targeting as it does the financial players who already know it so well. Then there was also what he somewhat snippily terms the push furor. Marimba, Pointcast etc. started to talk about push, meaning the idea of publishing web pages from web servers to web clients, a pernicious nonsense which had begun to confuse Tibco customers as to what push was, for Tibco, with the glacial confidence of professors, says it is the only firm that does true 100% push. The Reuters division has in fact been able to line up a lot of the first generation of push players by offering to provide Tib as the plumbing underneath what they do, and though Marimba and Pointcast have as yet not signed up Bowles is almost maddeningly confident that they will do so very soon – Within the push arena we primarily sell plumbing and they primarily sell applications – in the long run that’s a marriage made in heaven. They’re just customers who haven’t signed up yet. So if Tibco is prepared to go and sort out these confused push kids, it’s just as nobly prepared to set Corba types on the right path to (you guessed it, message oriented) nirvana. Tibco has been a contributing member of the Object Management Group since its outset, and has had an Orb of sorts out since 1991, but now has a Corba 2.0 version product, TIB/ObjectBus, available. But don’t take that to mean Tibco likes the Orb world all that much. For, like push, this is a drollery it will tolerate so long as it gets its foot in the door to sell the real stuff. I would be happy being just a participant in the Corba marketplace, just placing or showing rather than winning, as long as it gives us a chance to expose our message oriented middleware to Corba-oriented buyers, Bowles says. He is quite happy with the suggestion that he will sell his Orb to a customer, perhaps a chief information officer told to get distributed computing up and running by his CEO who knows about Orbs from the press but who maybe hasn’t heard of Tibco yet, and when the project screws up six months down the line sell him the real goodies. That at least seems the implication of a statement like Once the customer has purchased an Orb, and runs into problems with information dissemination or distribution, and then into problems of deployment, especially if he needs 24 by 7 support and has to buy 3,000 copies, middleware – especially if it’s to do with instant news, the notification of change and especially if they require a degree of customization – becomes ideal. So to compete with Tibco, a fellow Orb vendor simply has to say, Don’t buy from those guys, because they talk Orb but really want to sell you a much higher-ticket item? Bowles pauses for a good ten seconds, then suggests that a fellow Orb player would only suggest that if they were themselves frightened of being unable to deliver six months down the line and hence lose the customer anyway.

Spreading the real gospel

Is this a slightly disingenuous approach to both push and Orb work, since Tibco is basically saying both are fairly misguided attempts at an answer it already found ten years ago, and that it is dabbling in both markets simply as a way of spreading the real gospel? One has to say yes, though all being fair in love and the software industry there’s nothing wrong in such a tack at all. Tibco is still really just a systems integrator that sells blindingly good message oriented middleware technology, but which still really has no idea about marketing. It would make this particular reporter happier with its message, for instance, if it stopped using terms like push and Orb as if it meant them, and not just as loss leaders to get one to think messaging. But at the same time what Tibco says in a way doesn’t really mean much. It’s a wholly-owned subsidiary of Reuters, and despite the ink about Cisco Systems getting some undisclosed minority share in March it probably always will be. How separate an identity can it ever really have? What’s the incentive for its officers to really beat the streets to gain that much mindshare, since they are unlikely to ever get any stock options in any hot ticket IPO? It does most of its selling through VARs anyway, and has a very small direct sales force, and presumably most of its dollars (actually down 22% year on year according to Reuters fiscal 1996 figures in February, to 60m UKP, making one wonder if Bowles’s claim that it has been profitable since day one all that defensible) from its large ongoing bodyshopping contracts rather than shrink-wrap software sales. Tibco, like its estimable product, should maybe stick to being a plumber, not a player.