These are the old pages from the weblog as they were published at Cornell. Visit www.allthingsdistributed.com for up-to-date entries.

December 19, 2003

Is this the state of SOA research?

I haven't posted anything about the technical program of Service Oriented Computing Conference program because I can't make my mind up why I am so disappointed with it. Maybe my initial enthusiasm for meeting others that are doing research in the services area drove up my expectations to unreasonable heights. And with 180 submissions you would expect that the accepted 20% is good quality and interesting, even if it is outside your focus area. Unfortunately most of the papers stop at the conceptual level. I have seen many 'designs' in the past days, but hardly any implementation. Even to the point that the more 'systems' like presentations end with something like 'I will implement this if I can get any funding for this idea'. I was shocked, because some of this work is only valid if you can actually build it. None of my students would even dare to suggest writing a paper, without implementation and experimentation.

I don't mind a few conceptual papers in a conference. Especially if they are really visionary or take an extremist approach to solve a particular problem. But most of these papers are in essence research proposals, not reports on results. I have to say that a positive exception to this are the IBM Research folks I met; Sanjiva Weerawarana, Stefan Tai and Frank Leyman, clearly already have the kind of experiences that make an exchange with them interesting and challenging.

Maybe this conference was too early (and my expectations too high) and there are just no results yet to report yet, and we will have  wait for another year or two to get the quality. ICSOC 2004 will be in New York City, probably at the Columbia campus. I will certainly report on the state of my research at this conference again, I hope that others actually building SOA software will also do so.

Posted by Werner Vogels at December 19, 2003 04:09 AM
TrackBacks

No Original SOA Research?
Excerpt: Werner is disappointed by the state of SOA research. I didn't attend the ICSOC but I can imagine how researchers are just repeating 20 years old stuff in angle brackets. No surprise SOA is a vague term - even Don is refusing to use it. From the compu...
Weblog: Radovan Janecek: Nothing Impersonal
Tracked: December 20, 2003 02:23 AM
Comments

Maybe the people that build SOA products are just too busy to report the results... I somehow missed the conference. If you have any details about the next one I would appreciate them. Systinet would certainly have interesting topics to contribute!

Posted by: Radovan Janecek on December 19, 2003 05:45 AM

He, I am busy too... :-)

I understand that the product people are busy. I didn't expect you guys to show up (althought I think it would be very, very good for the research world to get a reality check).

But I am wondering whether in the research labs and in academia people are just rehashing stuf they have already done in the past (e.g. workflow, mobile services, personalization) and are now just redoing these in the context of service architectures. So none of this is original SOA research.

Posted by: Werner on December 19, 2003 06:14 AM

Theory is clean and evolving in itself. Implementation is clumsy and could even wreck expectations.

Posted by: Vishal Shah on December 19, 2003 04:19 PM

I think that most of SOA stuff is just a rehash. Some of the 'bracketing' is OK and some is hurting (like some of the transactions specs). Anyway, it seems to me that SOA innovation is more driven by product vendors than academia. We do original research it is not public yet except few vague posts on my blog.

Posted by: Radovan Janecek on December 20, 2003 01:30 AM

To Vishal: Absolutely disagree. Especially in the IT context.

Posted by: Radovan Janecek on December 20, 2003 01:39 AM

Werner:

I tend to disagree with your feedback on the conference. We may not have attended the same sessions, but I felt I got a lot of (real not hyped) information from the academic space and also from IBM. A lot of this information can go directly into building our product. I also felt that the academic community had its feet on the ground when it came to working on "coordination".

SOA is very diverse, quite complex and mostly unchartered. I got a feel for all that from the conference. There is of course some level of rehashing but anyone who beleives that SOA is just the next generation of application servers might be missing the point.

JJ-

Posted by: Jean-Jacques Dubray on December 22, 2003 10:08 AM

> Absolutely disagree. Especially in the IT context.

IT? How long has SOA been here? The theories might repeat, even collide, that's what happens in every field and that's good news. I don't see anything wrong in conceptual papers with little or no implementation. Otherwise, you confine your research abilities.

Posted by: Vishal V. Shah on December 24, 2003 12:42 AM