Meaningful debates about the viability of Web services are beginning to emerge
by Justin Kestelyn |
As the adage goes, if you can "make it" in New York, you can make it anywhere. Based on the discussions at XML DevCon Spring 2001, the XML developer conference that convened footsteps away from Times Square last April, the Web services paradigm is a long way from measuring up to either standard.
The Web services concept more or less entered the public IT consciousness in 1999, when Hewlett-Packard announced its E-Speak initiative as enterprise computing services chief Ann Livermore grandly proclaimed "the second chapter" of the Internet. Since then, the other powers-that-be - IBM, Microsoft, Sun Microsystems, and Oracle, as well as a host of smaller vendors - have coalesced behind the protocols underlying the Web services paradigm with rare alacrity. Consequently, the industry has been fire-hosed with news about Web services partnerships and vendor initiatives - a public-relations onslaught that recently culminated with Microsoft's .NET announcements, which, whatever you think about that controversial company, clearly demonstrate a boldness that is lacking in many of its competitors.
The premise involved is that the addition of a business interoperability layer over HTTP will let developers craft new systems by aggregating self-describing objects, whether small business components or arbitrarily complex ones, that live outside the firewall.
Voices of Reason
Some of the keynote speakers at XML DevCon publicly extolled the Web services framework as the end-all of application development and B2B e-commerce, the means by which the Internet will fulfill its ultimate promise. (Similarly, in a different venue, one IBM evangelist went on record with InformationWeek to proclaim that Web services would be "even more impactful" than the original Web itself.)
Privately, however, some of the same speakers raised serious doubts about the ability of the standards bodies and vendors involved to deliver the goods; at least, not anytime soon. And others - some speaking on keynote panels, some in small sessions - voiced doubts about the battle-readiness of key Web services protocols such as the Microsoft-spearheaded Universal Description, Discovery, and Integration (UDDI), for Web service registration and discovery; Simple Object Access Protocol (SOAP), designed for messaging transport; and ebXML, for core business process "vocabulary." For example:
- David Orchard of Web services consultancy Jamcracker noted several gaps in the currently proposed Web services stack, including a lack of adequate security for message routing via SOAP, and raised questions about UDDI's ability to scale.
- EAI expert JP Morgenthal made the excellent point that because nearly all complex B2B processes inevitably require some kind of human decision-making, the goal of a completely automated "business Web" is, at this point, far-fetched (but raises some interesting wireless implications).
- Norbert Mikula, CTO of enterprise portal company DataChannel, questioned whether anyone had given serious thought to the critical, but hardly glamorous, issue of service-level agreements (SLAs). How exactly would SLAs work on the automated business Web, particularly when mission-critical processes are involved?
- Finally, several speakers emphasized the fact that lacking completely open, ubiquitous protocols, Web services (as currently envisioned) will utterly fail. Thus far, SOAP is the closest thing we have to a standard, and even that outcome isn't assured.
Take Comfort
The good news is that the hype, like fog on a spring morning, is starting to clear. As we shuffle toward consensus, we can take comfort in the fact that the right questions are being asked, and without those, there will be no answers.
|
|
|
|
|