With Microsoft’s push to the cloud people are publicly wondering about the amount of R&D investments being made in “traditional” on-premiss server software. Microsoft is making a big bet on cloud computing and they *will* succeed. Looking at BizTalk Server specifically, and Microsoft integration middleware in general I can tell you this:
In the past couple of years, Microsoft faced three major hurdles in the adoption of BizTalk Server:
1. Low latency challenges. This prohibited BizTalk being used in people facing applications.
2. .Net developers
3. The “two workflow engines” story
The first is commonly known. Until BizTalk 2010 you had to go into the Management Database to tweak some settings in order to accomplish lower latency scenarios. It still meant a minimum of 50ms for every persistence point, but for most scenarios it was okay. This could however not be combined with high throughput scenarios. Until BizTalk 2010, where these tuning settings can now be performed at the host and host instance level providing for combinations of low latency and high throughput solutions. But still…. every message goes at least once to the MessagBox database…
The second is probably less commonly known, but it is a fact that the “.Net developer” is the greatest BizTalk Server competitor. Why buy this stuff if you can build it yourself?
The third is also more and more known; why do we have two separate technologies and engines for handling human (WF) and system workflows (XLANG).
Together with the push to the cloud, this has resulted in some smart people at Microsoft coming up with a new architecture to make all of the above possible and satisfy everyone: one integration middleware architecture that enables development, deployment and management of integration solutions with one set of tools and which can be deployed in the Cloud and on-premises (Server or Appliance).
Watch for some cool stuff being released in the coming period. Some of it already has been released, such as this week’s CTP release of composite apps in Azure.
It will probably take them a couple of years until this all proves to work out well, but I’m confident they’ll manage. And for us integration specialists it will become better and better, because integration won’t go away and the architecture and tools will help us architect, design and build better integration solutions and at the same time bring new possibilities thanks to the cloud. This will be a real paradigm shift!
Cheers! Gijs
indaField 