Sticky Strands of Web Services Uncertainty
Page 1 of 2
So, just who is the "A" team for Web services architecture? Is it Microsoft's .NET, bustling with newness and opportunity? Or is it the time-tested Sun Microsystems-led Java language?
For the uninitiated, Web services are pieces of software that can be integrated and run on distributed computers throughout the Internet or on a network. Major vendors, including IBM, Microsoft and Sun, are investing heavily in the sector in the hopes that their technologies become widely accepted. Many research firms estimate the market for Web services will be worth billions by 2006.
A recent study from Evans Data Corporation's North American Developer Survey series found that the battle for Web services technical standards is as about as even as it can get. In a survey of more than 600 developers, the research firm found that while 40 percent of developers are now creating apps for .NET, 63 percent will develop applications using the .NET framework a year from now. In turn, while 51 percent of developers surveyed currently write with Java, some 61 percent said they planned to take it up in 2003.
Dan Hay, a product manager for Visual Studio .Net at Microsoft, told internetnews.com that although Microsoft doesn't necessarily agree with Evans Data's methodologies, that the figures are consistent with what he has heard in the industry.
David Harrah, marketing manager for Sun software, told internetnews.com that the findings aren't a surprise at all, that they are consistent with what he's heard on his end.
"That [.NET] is what we're working against," Harrah said.
Are Microsoft and Sun headed for pitched battle?
ZapThink analysts Jason Bloomberg and Ronald Schmelzer took a stab at interpreting and extrapolating on the Evans Data for internetnews.com. Neither is sure that choice of platform is a major issue.
Bloomberg told internetnews.com the choice of a platform gradually fade in importance as Web services evolve to be full, "coarse-grained" business services.
"Software development isn't going away, to be sure, but the choice of platform or language will become more of a "right tool for the job" question, where developers will pick and choose whatever language is appropriate for the task at hand," Bloomberg said. "In fact, the Evans data hints at this -- after all, with the two key numbers both well over 50%, the number of developers who are planning on using both .NET and J2EE must also be going up quite rapidly. And why is that? Partly because they want to use the right tool for the job."
Fellow analyst Schmelzer takes the concept further. Schmelzer said the survey indicates that developers are confused and, more interestingly, he theorized that .NET could eventually cannibalize (eek!) Java.
"With 63% of developers planning to implement .NET and 61% of developers planning to implement Java, we can notice that there is an increasing trend towards overlapping development. Basically, Microsoft has done an excellent job of telling developers that they can develop on top of .NET in *addition* to developing on top of Java. Why rip and replace when you can "embrace and extend"?
"The challenge with this scenario is that while organizations can double-team on .NET and Java, the reality is that developers cannot. A developer has to make a choice as to which platform they would like to develop on as the technologies and skills required for .NET are very different from the skills and technologies for J2EE. However, organizations aren't as financially wealthy as they were before. They are operating with far fewer developers and dollars, so they really can't afford to be duplicating their developer efforts. It seems that the above behavior is more a reflection of uncertainty as to where the development activity will happen rather than a choice. Organizations would rather choose one platform rather than many, but the evidence of the survey, if we can bank on its reliability, shows that they are increasingly not making a platform choice -- out of confusion."
"Part of this confusion is over what Web Services and 'interoperability' really means from a practical perspective. Both the .NET and J2EE camps will tell you that Web Services are an interface or abstraction layer, so that Web Services developed with .NET should be able to interoperate with Web services developed in J2EE. So, this begs the question, why do organizations have to support both platforms? An organization can afford to choose one platform and be 'guaranteed' that it will interoperate with other platforms, within or outside the company. This is a benefit to Microsoft, since IT organizations can now be able to cope with singular Microsoft servers co-existing peacefully with a bunch of J2EE servers. However, this also goes to the point of development -- organizations will as a result migrate to platforms with greater ease of use and lower total cost of ownership -- an area where Microsoft has already been strong. We should see that if Microsoft lives up to its promise, it will continue to eat away at the J2EE market share and we'll see the Java numbers start to reverse while the Microsoft numbers increase. This overlap percentage is a temporary blip -- at least for the next 2 years."
Web services has generated a lot of buzz, sure, but there are problems that need to be ironed out. See Page 2.