Some wise words from three influential personalities in the web services field. Sun's Jonathan Schwartz and BEA's Adam Bosworth are quoted from an article in the new and very useful Enterprise Architect magazine, The Making of the Services-Oriented Enterprise, by Steve Gillmor. Access to the site is by free registration. The others are from an interview with BEA's Scott Dietzen by Michael Singer of siliconvalley.internet.com.
Jonathan Schwartz: "The reality is that the game has moved beyond 'What's your app server?' The game has moved to 'What's your directory server, and what's your identity strategy?' The most fundamental web service that exists in the world is logging in, and the single biggest integration challenge every company faces is stitching together redundant authentication systems."
Scott Dietzen: "The reality is that there has always been five times as many business programmers as there are system programmers. There are many more Cobalt, VB power builders developers than there are CORBA, J2EE, distributed object developers because business programming is easier than systems programming."
Adam Bosworth: "There's a higher level of loose coupling that's required in the enterprise as well, and that's a loose coupling where at the time you build apps, you don't know what other app is going to be on the other side. This is loose coupling in the large, loosely coupled around which applications are actually talking to each other and not just how they talk to each other."
I was going to stitch these comments into a longer posting, but that task has been lying on my to-do list for long enough already, so here they are in the raw. Scott's interview also has some good stuff about the state of play with various web services standards and where they fit, plus a nugget of impromptu market research: "... all of [our top users] claimed they had transactional web services in production and 60 to 70 percent had transactional web services running between .NET and WebLogic in production."
Coming back though to the quotes I chose to include above, here are the takeaways I wanted to highlight:
Digital ID is now prime time. If you don't develop a distributed authentication and access infrastructure that gives your users single sign-on to web services, you're going to be toast.
Development is stratifying. Systems programmers are becoming responsible for application infrastructure. That frees up business programmers to focus on business infrastructure. Business users are becoming developers too, with increased access to information systems being delivered to desktop tools like Excel.
Loose coupling is fundamental. If you don't design it into business systems from the ground up, you're just tinkering at the periphery. That doesn't mean you have to rip-and-replace your entire infrastructure, but it does mean your systems developers are going to have to create some great middleware and your business guys are going to have rethink the way they plan and operate business processes.
Microsoft may acquire a web services management company or product, Steve Ballmer hinted in a conference presentation yesterday. According to Microsoft's local paper, the Seattle Post-Intelligencer:
"During an onstage interview at the Gartner Symposium for information-technology professionals in Orlando, Fla., [Microsoft CEO] Ballmer said to 'stand by for news' about an acquisition. His comment came during a portion of the interview dealing with software development tools and XML Web services, small pieces of software that let programs on disparate systems interact.
"Ballmer made the statement in the course of underscoring the importance of letting software developers determine, at the time a program is created, how it will be 'managed,' or centrally controlled and monitored over a network."
The comments suggest that Microsoft is eyeing a management startup that has emphasized building in management logic right from the start of web services development. One of the strongest potential candidates is AmberPoint, which in February this year announced a tie-up with Microsoft to build "a combination of tools and management technology intended to help customers manage distributed applications based on the Microsoft .NET Framework." But AmberPoint is also close to other vendors including IBM and Computer Associates, and already integrates its products with BEA development tools, so it isn't a foregone conclusion.
Another strong candidate is Actional, which, like AmberPoint, is a member of Microsoft's Visual Studio Industry Partner program, and already offers close integration of its products with BEA development tools. Earlier this month, Actional completed a deal to sell off its integration adapter business to iWay, which could be interpreted as clearing the decks ahead of an acquisition of its core business.
Other potential candidates include Blue Titan, Confluent Software, Digital Evolution and Flamenco Networks, all of whom have declared partnerships with BEA, whose promiscuous willingness to partner with just about everyone in the web services management field means they can all boast of their experience without really having any sense of commitment to the relationship.
If Microsoft does make an acquisition, it will be following in the footsteps of webMethods, Computer Associates and Hewlett-Packard, each of whom have recently acquired web services management companies. But perhaps Ballmer was just having fun with us. Whether he made a genuine slip of the tongue or whether his comment was a carefully calculated hint remains open to debate. "With Steve Ballmer, you never really know, because he has been prone to do both of those types of things," Michael Gartenberg of Jupiter Research told the Post-Intelligencer. Later on, The Seattle Timesreported that Bill Gates had downplayed the comments: "I don't think he meant to suggest there was something imminent," Gates said about Ballmer's comments. "Every year we do buy some additional software companies. There's nothing in the works that he was trying to hint at there."
posted by Phil Wainewright 1:13 AM (GMT) | comments | link
Tuesday, October 21, 2003
Who dares wins
How do you handle disruptive change in your established market? Last week's flurry of announcements and acquisitions demonstrated several different approaches some of them flawed, I would contend. See below what I think of, respectively, BEA, Seebeyond, Iona, webMethods, EMC and Siebel.
Raynor and Christensen's article last week looks at how Sony rode the rise of transistor technology in the 1950s to crack the prevailing market dominance of manufacturers of vacuum tube-based radios and TVs, such as RCA. Why is this at all relevant to software vendors in the 2000s? Because, say the authors, "it is stunning to see how the sins of the past so regularly visit the later generations of disrupted markets." The Sony story illustrates some key points:
Innovations compete best against nonconsumption "Because [Sony's] customers' reference point was having no television or radio at all, they were delighted with simple, crummy products."
Chasing value breeds false security When Sony began winning low-end customers, RCA and its peers were moving upmarket into color TV. "The vacuum tube companies' profit margins actually improved as they were disrupted."
Established distribution and sales channels aren't competitive Sony created a new channel, selling through chain stores and discount retailers. "Appliance stores that made most of their profits replacing burned-out vacuum tubes in the products they had sold ... couldn't make money selling solid-state televisions and radios."
It's a tall order surviving when faced with disruptive technological change. Here's how I think a selection of companies are doing:
BEA was the subject last week of a Loosely Coupled article analysing its SOA strategy, BEA stakes future on SOA adoption. It's a courageous strategy, but I think BEA has the problem of being caught between two disruptions. Yes, services-based development that focuses on assembling together existing functionality is the way of the future, and WebLogic Workshop is a great tool for doing that. But customers especially those with a bias towards Java are increasingly going to want to deploy to open source platforms like JBoss, and that's a disruption too far for BEA, which is counting on a healthy revenue stream from its proprietary application server implementation of J2EE.
SeeBeyond last week announced new versions of its application integration tools with enhanced web services support. According to a report in ADTmag.com, compay VP Alex Andrianopoulos "maintained that his company is ahead of competitors in employing standards, including the Web services Business Process Execution Language (BPEL)." But I'm afraid SeeBeyond has its eyes on the wrong competitors. It's no good staying ahead of other EAI vendors. The challenge is to catch up with the next generation of application assembly solutions.
webMethods took a courageous leap with a series of acquisitions, the most notable of which was of The Mind Electric. I think its embracing of TME's "web services fabric" technology gets the company where it needs to be. Certainly all its statements last week were designed to convey the message that webMethods has moved beyond the old EAI world. But that doesn't mean its troubles aren't over. To make the most of its acquisition, it now has to court a new customer base, with a new distribution and sales model. If it has acquired TME's fabric without taking on board its strategy of broad market distribution, it may still fail to capitalize on its highly imaginative initial leap.
Iona was somewhat unlucky to have its Artix launch overshadowed by webMethods' announcements. Here we have another imaginative leap into a next-generation technology offering. But Iona's go-to-market challenge is probably even more marked than webMethods'. It doesn't currently have the distribution and sales channels it will need to follow through on that announcement.
EMCannounced a huge $1.7 billion deal last week to buy content management giant Documentum. This is supposed to be a deal with lots of synergy, but frankly I don't get it. I never do get it when people say, well, we can't make money in our core business, so let's buy a business with a bigger profit margin. That doesn't fix the problem you started out with, it merely disguises it. Their sales guys are supposed to be selling to the same customers. Well, you'd have to find a lot of redundancy to restore margins just by consolidating sales. And what about the sales guys who sell EMC hardware to Documentum's competitors? Buying up channel partners is never in my view a good long-term recipe for enhancing your business performance.
Siebelbought Upshot. Well, well, well. I still hold to my gut feel that Siebel is an inveterate dinosaur. In my view, the logic of disruptive innovation is that its destiny is already sealed. The sales channel problem likely means that Siebel will never be able to adjust to the business model required to sell products like UpShot. I can't also help wondering about the coincidence of Upshot, one of whose main distinctions was its use of MS Office applications as offline clients to its software, signing a deal just a week before the launch of Office 2003, which effectively obsoletes that feature by building it into Office as a default for any XML-capable source. But as you can see, my mind is made up already. It will take a lot to convince me that Siebel can save itself after all.
OK, I'm off to see the launch of Office 2003 now. I wonder how Microsoft will do?
posted by Phil Wainewright 6:53 AM (GMT) | comments | link
Assembling on-demand services to automate business, commerce, and the sharing of knowledge