Where will the potential land mines in the Web services scenario be found?
One is the standardization process around security. If that doesn't proceed at a reasonable pace and if it were to get into a lot of industry arguments about this and that--I'm really hopeful that somehow we sort this thing out so we get less posturing and more working on the guts of these things. Security is a (subject) where all of the vendors have a responsibility to behave themselves. It is absolutely critical to the real use of Web services.
The second thing is that as people design additional Web standards, you have to have the right degree of sophistication and simplicity. It would be really easy to reinvent a complicated scheme that reproduces everything we have ever done in the past. And that would move it away from the elegance that has so far characterized a lot of the Web services work.
I think we are OK so far. But as we finish this process, we have to be wary of getting carried away and moving too far away from the simplicity of Web services.
Is there going to be a training hurdle here for IT developers?
The customers I have talked to understand this stuff very well at the developer level. You can't pick up a trade journal where they're not talking about Web services all the time. They hear a lot about it, and developers have reached a point where they can cut through a lot of the vendor hype. They are trying to make decisions now on things like Java versus .Net. I think developers are getting it. I'm pretty confident about that. The questions I'm getting are great questions, and I think developers are up to speed.
What questions are customers asking?
They're asking the pragmatic questions: What do you have now? Where is this going? Whom will you interoperate with? Give me the confidence that I can make an intelligent decision--that I can weigh the risks here--about when I can adopt this technology.
And we say: Start a pilot (program) somewhere. The safest way to do this is to get yourself educated. You will know about your own team, and you will know how these technologies work in your IT setup, which is different than anyone else's setup anywhere in the world. Until you do that, it may be difficult to do a risk assessment.
Are customers asking about the benefit of using (Sun Microsystems') Java compared to (Microsoft's) .Net for Web services?
We get a lot of questions about .Net and what it is. In many ways, .Net is just the next generation of Windows, with better ways to communicate to the outside world. But in many cases, it all boils down to Windows talking to Windows. So people may make that choice. But when you look at the server side of the world, that's where the information is going to be flowing, and that's where you get into the questions of security and other issues...Then you get down to the competition about development platform, the strategy that IBM has adopted versus Windows.
Will Web services mitigate the Java versus .Net decision in any way?
Well, it changes it a little bit. IBM is an immense supporter of Java. But for us to make some sort of statement that we will only let you build things that only talk to other customers using Java is inane. Nobody would want to do that.
It should be the case, though, that if I have built a Web service using Java on WebSphere, and you use .Net, you should be able to talk to us. I don't want to know you're using .Net, and you should be able to rebuild your application next week using Java, and I still don't need to know. It really comes down to less of an issue of Web services per se, than it does to making intelligent decisions about your platform where you're building things.
Why has it taken this long for Web services to catch on?
There was a lot of confusion when it first started, about what is the consumer (angle), and did it have anything to do with productivity applications? If I'm using a consumer application, I shouldn't know there are Web services behind it. I should just think it's a great application.
I think that a lot of the corporate-vendor posturing is decreasing on this. There is this feeling that we are moving into the second phase to where it will be hardening down to what Web services is. The next year will be fascinating in terms of what people can really do with Web services, and what they're using Web services for.
In what unexpected ways is Web services evolving?
In the Web services area, things have always happened faster than I predicted. I really did not think we would be into things like workflow and transactions until sometime next year. We will probably hit that four or five months earlier than I thought.
There is more of a psychological acceptance of the concept. Sometimes you get agreement faster than you thought you would. That has been my biggest surprise.
Have Web services progressed the way you expected?
A year ago September was very different from what anyone imagined it would be. And so in many ways--and appropriately so--it put the brakes on a lot of this technology development. Things got very quiet for a while.
But when we started this year, you could really feel a lot of pent-up energy around Web services because we had this long period, six months, where there had not been a lot of Web services announcements.
What remain the most important issues to resolve?
Security and completing the pieces of the security road map published last April that went to OASIS (Organization for the Advancement of Structured Information Standards) in June. Also, things like WS-Policy, WS-Trust (Web services standards proposed by IBM and Microsoft). That stuff is well understood. It's just a matter of knuckling down and doing the work.
Sounds like you're confident that agreement on a common way to do security is within reach. What other question marks are still out there?
The two question marks right now are around reliable messaging and systems management. A lot of the questions around reliable messaging are around things like, well, I'm in a company and I use (IBM's) MQSeries (middleware). That gives me reliable messaging, so why would I have to do anything else? So then you say, well how about (messaging) across the Internet? So you always keep returning to this question, saying what about this scenario, what about that scenario?
What we are really in the process of doing right now is creating a model that can say, all right, I can do this at the SOAP (Simple Object Access Protocol, one of four key Web services specifications) level, I can do this at the application level, but I also know how to take advantage of the underlying protocol if it is powerful enough to provide me some capabilities already. Then you won't have to duplicate effort. So that will be coming out. You will be seeing some things there.
Systems management, I think, is the last big thing in Web services. It's building on a lot of the experience we have had building on standardization efforts in the DMTF (Distributed Management Task Force).
In which areas of IT will Web services be appropriate?
Usually the issues center on things like security and speed of transactions and number of transactions. So you can start by thinking of things like Web services in situations like B2B (business-to-business) situations. You and I want to do business. I don't care how you build your system; I just want to do business with you.
The idea with Web services--like the Web itself--is that if we can get agreement on standards, and use things like XML (Extensible Markup Language) and the Internet, you are going to have the universal connectivity of the Internet. And, just as all Web browsers handle HTML (Hypertext Markup Language), all application servers now handle XML parsing and SOAP. You are going to get to that same level, where you will be able to say, If I do Web services, the person on the other end of the line is probably going to be able to talk to me. That's the breakthrough.
Will Web services spell the end for EDI (Electronic Data Interchange, an existing way to build systems to exchange data)?
You will get a lot of continued use of EDI. People have made real investments in EDI and it works just fine, and nobody is telling them they need to replace that. But if anything has come about in the last six or seven years, it's this increasing understanding that I have information that I need to put out there, but I don't know who the ultimate consumer is going to be, like a Web page. You don't know whom all of your potential customers might be. You don't want to start using technologies that will somehow put you at a disadvantage.
Click here to read full interview
Back to Vision Series