The guys I work for were sponsoring the Microsoft Real World SOA summit here in Aus. so I went along to represent the company. I sat in on a great talk from Kris Horricks from MS CSD. He went over Oslo and how that feeds into some of the recent developments coming out of Microsoft. For the first time I got a good look at the feature set of Dublin. At first I was gobsmacked. Asynchronous messaging, long running transactions, correlation, workflow, WCF messaging, tracking, persistence, dehydration/activation. Sounds like a feature list for BizTalk Server doesn’t it. This will all be packaged up in Windows Server 2008 as an application server role with interaction via MMC. The question in my mind was are we witnessing the end of BizTalk Server. I mean who would pay the licence fees for BizTalk when for a fraction of the price Windows Application Server can be had.
However there are distinctions between the new application server and the integration server we all know and love. E.g. adapters, edi functionality, advanced load balancing and throttling, RFID, to name a few… I can almost here the phone ringing at the office now with questions relating to BizTalk being a poor strategic investment in context of what’s happening with Dublin. I fought the same battles when .NET 3.0 came around.
At the moment I see them as complimentary technologies that do different things. If I want to develop a set of WF services to build composite apps then I’ll need a hosting service for that composite app. Windows Application Server(Dublin) is the natural selection. If I want to integrate with numerous legacy systems talking EDI etc tackling the more traditional integration scenarios then BizTalk Server is the natural selection. I can envisage scenarios where the two products would work nicely together.
That might be the case today but I wonder where we’ll be three years from now. Microsoft tell us both products have independent roadmaps but there is still a question in my mind.