Whither WSE?

Yesterday I did a workshop on “Best Practices for .NET Enterprise Architecture” at the Enterprise Architecture Summit 2004 in Palm Springs. One of the questions I asked the 30 people in my workshop was “How many of you are familiar with WSE?”. About 10 hands. Then, “How many of you are using or are planning to use WSE”? Zero hands. Whoa! Then, “How many of you are not going to use it because of something the Indigo team has told you?” The same ten hands. Which pretty much confirms what I heard at TechEd, which was customers telling me that the Indigo team had told them to “hold off“ on WSE.

Great. Organizations are going to hold off on implementing SOA using .NET today because of some vague promises of a future technology. Please, somebody just shoot me now.

8 Comments

  • Is that Pleas somebody shoot me now?



    Keith, I was in an architecture meeting this morning discussing a .Net/Oracle Financials interface where a decision was made to use oracle tables and triggers as the messaging mechanism. SOA and web services are far removed from the mindset of the average architect/developer.



    There are a couple things that will hurt SOA from what I've seen so far but the biggest is this... SOA requires expertise and cooperation on both sides of the table. One enthusiastic person is not enough to convince a room full of people (Oracle and .Net) that services should be considered.



    Also, imo WSE will hardly be used at all in production code. Plenty of code samples will exist but actual production implementation? People just are not taking web services seriously yet. I would have had an easier time this morning selling an inappropriate use of MSMQ. Interoperability scares people into making safe time tested choices.



  • I'll give you some comfort... I'm part of a start-up company and we are going the SOA route. Currently, the only way we can implement our software is by using WSE 2.0 and extending it beyond the specifications currently out there that weren't implemented in WSE.



    I have to say, once I got WSE 2.0 last week, our entire architecture clicked and we're moving fast ahead into production code. We can't wait for Indigo to make things better for us when we need to hit the market in time.

  • I have a hard time seeing how people got that impression.

  • Without getting into the technical merits of WSE / Indigo etc etc, this has unfortunately been something Microsoft has been guilty of for a while now. It used to be that betas were private, the vast majority of the population were not aware of what was coming next. Now we have insight into the the next 2 versions of .NET. Why should I use WSE when Indigo is "just around the corner". Of course we won't see Indigo in full bloom for at least another 2 years, but that point just doesn't seem to connect with anyone. Not sure what the answer is.

  • Box says that in the PDC webcast from November. I've heard a number of times that WSE 2.0 will migrate "nicely" to Indigo. If I was at Keiths talk I would have stuck up for WSE.

  • WSE 2.0 is a Microsoft supported product. It's not a good choice for long term development (say more than 2 years), because the WSE release cycles are too short. But nonetheless we're using it as part of our HL7 messaging solution (very much a PRODUCTION system), because it fits the bill. Also note that WSE 2.0 to me is (practically) the first .NET SDK that allows me to host webservices in my own application.

  • Keith,



    Equating WSE with service orientation or SOA is a huge leap.



    You don't need WSE to build service-oriented systems on our platform. ASMX is an excellent solution and waiting for WSE or Indigo or anything else seems odd.



    WSE does add some advanced security support that is important if you need something beyond transport security, but to then say that all services need WSE is a bit of a stretch.



    DB





  • Keith, we are already using (in production) a messaging subsystem using WSE 1.0 and we are currently making a new version based on 2.0 architecture and capabilities.



    We know that all (or almost) of the code we write for WSE 2.0 will be not portable to Indigo but we need this now and i think that the benefits and experience we gained (and will gain) of implementing this now are invaluable.



    Best regards,



    Andres Vettori

    MCSE/MCSD/MCT

Comments have been disabled for this content.