This post originated from an RSS feed registered with .NET Buzz by Dave Bettin. Original Post: Q&A for Services in 2005 Feed Title: Dave Bettin on Services Feed URL: http://www.davebettin.com/me/rss.aspx Feed Description: Bettin' on .Net Service Development Latest .NET Buzz Posts Latest .NET Buzz Posts by Dave Bettin Latest Posts From Dave Bettin on Services

My predictions for services in 2005 in sort of a Q&A format. Actually, now that I think about it maybe these aren’t my predictions for 2005 but more along the line of my wishes for the upcoming year.

 The Question: Will we see additional spec(s) from Microsoft and company released to the W3C or Oasis bodies in 2005?

The Prediction:    WS-Policy, WS-ReliableMessaging, and WS-Eventing will be released to the wild. WS-Policy, being the most important spec yet to hit the standards parade, will be submitted to the W3C to align with SOAP, WSDL, and WS-Addressing. (In the near future, it will be vary rare to introduce services without using these four specifications in some fashion.) WS-Policy will probably see the same sort of questioning as seen with WS-Addressing’s introduction to the W3C, but  will prevail nonetheless. WS-ReliableMessaging and WS-Eventing will be submitted the Oasis group and we will see the merger of  WS-Reliability and WS-Notification respectively cause it’s the correct move for the industry. Right? yes.

The Question:  I understand WS-Addressing was submitted to the W3C last year, but I have been lurking on the mailing list and I am starting to get the feeling that it won’t see the light of the day.   Will WS-Addressing be forever stuck in standards hell?

The Prediction:  WS-addressing is too important for the service stack to be stuck anywhere ,whether that be Microsoft’s specification vault or in the hands of the W3C. Although, right now we are feeling the aftershocks of different vendors/individuals interpretations of the spec; WS-Addressing will see the surface at the W3C and will become a significant force behind the growth of advanced services this year.

The Question:  Service performance on the wire is an absolute critical success factor in some many of my distributed apps but when large sets of data are encountered the performance of my services become akin to a snail’s pace.  Will 2005 be the year that the industry can abate the angle bracket bloat?

The Prediction:  I believe the Elephant in the Living Room will finally be talked about in much a broader forum this year and we will start to see creative/innovative solutions for this issue. Vendors will release their respective implementations of Binary XML and realize, while more performant, we lose a lot of  the inherent transparency benefits of XML on the wire and will start to heavily investigate streaming and/or chunking of XML messages as an alternative solution to the performance problem. Whether that be WS-Enumeration or something akin to Indigo’s message streaming model.. I am not sure. Will the perf problem be reduced? Yes. Will it be at level that is acceptable? Maybe.  

The Question:  It has been touted by service wonks in the past that my service should able to just work with that other service without any major surgery. Unfortunately,  that is not case and often requires long hours of surgery to make it work.  Is there any hope it will finally just work this year?

The Prediction: Yes, I do believe we will be at a point where different service stacks will play nicely with each other and just work.  But not because of interop workshops  but because of the efforts behind the WS-I group.  This will be the year that WS-I’s work will gain a lot of visibility (the WS-I advocates program launched late last year is a wonderful precursor to this enhanced level of visibility) and traction in providing constraints for the interoperation of services that just work.  These constraints or profiles provide a single interpretation of the sometime vague specifications and enable implementations to comply to WS-I profiles and test cases.  WS-I and WS-Policy will become a powerful combo in enabling services that interoperate dynamically and just work.

The Question: Will Mark Baker continue his tradition of dangling the failure of services. .err.. red herring in the faces of WS-Believers?  Note: I felt really compelled to put this question here because I get a nice chuckle when I see XMethods used as a measuring stick for the pervasiveness of  soap based services.

The Prediction: Of course.