Category: Development

And just after my post on if SOA does exist I see this article on the benefits of SOA.

The company hasn’t calculated potential cost savings from the shift to an SOA-based application development model. But Poulter noted that using Web services to integrate applications over the past two years has been 60% to 70% cheaper than traditional hard-coded integration.

I don’t understand how that can be. Maybe I’m stupid, but I really don’t understand how using something like webservices will save you 60%-70% over anything else. What was the alternative? Assembler? 😉

Clemens Vasters has a post where he questions the existence of SOA.

What’s different now is that it is easier, cheaper and likely more productive to create bridges between systems.

He confirms my suspicions. The tools makes it easier, but it’s just not a completely new architecture paradigm. And you bridge systems only when you need it, because it adds complexity.

Update: Roy Osherove joins in here. Straight on. 🙂

Trond Andersen has a post on Integration Severs here. He’s not sure if the way they work is a good or bad thing, and I’m not sure either. I like to stay in OO land as much as possible.

I’m not really sure when an Integration Server would be a good solution. It seems to me it is on it’s way to become a bit over used. With adapters to legacy systems I can see the need, and it might reduce the time it takes to integrate these with new solutions.

In most cases it will be just as easy to set up systems for direct communication. In some cases you could argue a Integartion Server provides loose coupling, but I honestly don’t see how it makes looser coupling just because you got JMS/WS/XML going over the wire. Just another thing I should have a peek at. 😉