More from Amy Wohl on web services and ASPs, this time in a considered article in her weekly email newsletter. I think she's right that the xSP business model now seems to have settled down and thank goodness for that considering all the blind alleys some of the early practitioners ended up in but her final comment has a sting in the tail for anyone who thought they could relax now: "Web Services will offer both infrastructure and functional xSPs the opportunity to rethink their offerings and to partner in new and interesting ways. That may change the market more profoundly than anything which has happened so far." Amen to that.
posted by Phil Wainewright 6:00 AM (GMT) | comments | link
Wednesday, March 27, 2002
Tibco's application-centric view of web services integration
Already the leading vendor in messaging middleware, Tibco believes the advent of web services has opened up new horizons of opportunity that will launch the company into the top league of mainstream software vendors. That was the message I took away from a lunchtime briefing earlier today with Alun Baker, managing director of Tibco UK. "Web services will drive the integration market exponentially over the next few years," he said. "It will take integration into the mid-tier space and [into] the slow-moving verticals that have been technology laggards." Or, as Geoffrey Moore would have put it, it will move Tibco out of the bowling alley and into the tornado. The company sees itself as well placed because, unlike many other middleware vendors whose intellectual property is concentrated in adaptors that convert low-level messaging, Tibco feels it has invested more in higher-level business process integration. However the company's thinking seemed to me still too firmly rooted in a mindset that sees software in terms of discrete applications that are built first and then integrated afterwards. So it only looks for competitors among the ranks of other integration vendors. But if you build applications using a web servcies architecture, what is the difference between an integration vendor and an application development vendor? From a business process point of view, I'm not sure that there is one any more. Maybe Tibco will have to fight on more fronts than its management expects.
posted by Phil Wainewright 9:56 AM (GMT) | comments | link
How one ASP sees the migration to web services
I had very welcome confirmation that it's not just me whistling in the dark here when I met yesterday morning with Eamus Halpin. He is CEO of iRevolution, an LSE-listed British company that has five years' experience of the ASP business and also still has a substantial solutions provider offering, both serving the small and mid-sized business market. Eamus has a far-sighted vision of where ASP, software-as-a-service and .Net-based web services are all leading, and he's just published a white paper (PDF, 332k) in which he's written it all down. One message that really stands out is that thin-client and browser-only solutions aren't sufficient. Eamus makes a very powerful case, based on iRevolution's substantial past experience of delivering line-of-business solutions using the hosted model, that many applications will have to be componentised in a way that allows them to run disconnected on the client device, even if they are remotely managed. This chimes in very neatly with the conclusions of the Yankee Group report I mentioned a couple of weeks ago. The white paper is well worth reading; I recommend it, especially to anyone who is trying to work out a roadmap that takes them beyond the limitations of present-day ASP models. Here's a taster (and no, I didn't prompt him to use my favourite phrase): "The software service solutions that are coming will be services that are designed from the ground up to be deployed as web oriented (or remotely deployed) services. These will be the true 'one to many' solutions; highly scalable, loosely coupled and richly functional."
posted by Phil Wainewright 2:50 AM (GMT) | comments | link
Tuesday, March 26, 2002
Online applications and enterprise software vendors don't mix
I was doing a final fact-check on my column this morning for ASPnews when I realized I would have to do a re-write. The SEC has stepped in with an enquiry that effectively stymies Network Associates' attempt to re-acquire its McAfee.com spinoff very fortunately, in my view. For a full blow-by-blow account, read the column. You'll find that I've used the occasion to climb back on my current hobby-horse, this time at Oracle's expense: "Anyone who thinks the ASP model is all about outsourcing conventional applications just hasn't been paying attention for the past three years. You either do it with a completely new approach to delivering software functionality, or you fail."
posted by Phil Wainewright 1:39 PM (GMT) | comments | link
Monday, March 25, 2002
Why web services need ASPs, and vice-versa
Last week a column by John Dvorak in PC Magazine likened web services to the ASP model and dismissed both as crackpot. Then on Friday (Mar 22), Amy Wohl commented in her weblog that "the ASP model and Web Services arenít the same thing". Now I know for a fact that Amy Wohl has been following the ASP scene for pretty much as long as I have, and she's right about many, many things in the business. While John Dvorak is a great columnist who keeps his finger on the pulse of the tech industry. But in this instance, I'm sorry to say they're both wrong. Here's why:
John says the notion that users would rent software on an as-needed basis "was as crazy when it was first proposed as it is today". Not true. What's crazy is the notion that users will rent traditional application packages on an as-needed basis. ASPs failed because they were trying to rent the wrong kind of software
You can't rent a package like Word feature-by-feature that always was a ridiculous notion. But you can rent web services component-by-component. That's the entire premise on which major chunks of web services architecture, such as UDDI, have been conceived. Web services were designed to be rented on demand.
Amy says ASPs are struggling with how to handle applications built to the web services model, but I think she's looking at that issue through the wrong end of the telescope. The trouble with renting web services on demand is that today the standards, platforms and expertise for doing so barely exist. That's a problem for everyone, not just for ASPs but it's a problem they understand better than everyone else, because they already know more than any other industry players about delivering software as a commercial service. On-demand web services will build on the technical and intellectual legacy of the ASP industry.
John advances a strong case that the formation of bodies like the Web Services Interoperability (WS-I) group with its "laundry list of drinking-club members" is a clear sign that "nobody knows what the heck is really going on." He may recall that the ASP industry had a consortium of its own, and I readily confess that I know exactly what he means. These consortia of vested interests are a concerted attempt to persuade the world that nothing revolutionary is happening here. They subverted the ASP movement until everyone believed it was all about outsourcing their failed software applications. Now they're attempting to persuade the world that web services are simply a new way of fixing all the in-house integration problems created by monolithic implementations of their enterprise application . Defining web services as a component of legacy software installations is a deliberate denial of their on-demand destiny.
John's killer argument comes down to money "an indication that this concept of software distribution will result in more profits than the shrink-wrapped model. Otherwise, what's the point?" I absolutely agree. But profits for whom? John and Amy have both been around long enough to remember the last time a new form of computing swept away the bloated, inefficient dinosaurs that previously ruled the landscape. A whole generation of technology vendors were wiped out by newcomers who used PC technology to deliver more effective computing to corporate America (and indeed the rest of the world) at a lower price. Exactly the same thing is going to happen when web services get going. On-demand web services are the next generation of computing.
This announcement from Sonic Software is the first out of this week's JavaOne conference to catch my eye. I have always had a soft spot for Sonic's Java-based message queuing middleware. Now the company has unveiled what it calls an 'Enterprise Service Bus', which is an interesting choice of terminology:
It's the first time I've seen the concept of a bus used in the context of services (a bus was originally the channel used for exchanging commands between components inside a computer, so applying the concept to service components is a metaphorical jump)
Tibco, the vendor that counts as the gorilla of the messaging middleware sector, was the first company to start conjuring metaphors for software messaging out of the bus concept, when it popularised its implementation of what it called 'the information bus' (hence its name, derived from the initial letters of that phrase). So I imagine we are supposed to infer that Sonic is now very consciously going after Tibco. I am having lunch with Tibco on Wednesday so I will be able to ask how they feel about that.