One year on: zero progress for Enterprise REST as Java races backwards

Just under a year ago I blogged about how REST and Sun had put Enterprise IT back five years so I thought it was about time to update that view and see what has happened in the last 12 months in the enterprise integration and governance space.

So on the REST front we've seen.... ummm.... struggling here.

Lets be clear, I'm talking here about REST as an enterprise integration approach, not as a way of exposing a Web API for content aggregation but as a functional integration approach for enterprises. Something to replace the "fundamentally flawed" WS-* that REST is so much better than. So what is the progress this year? Zip, zero, nada. Yup a few minor tweaks into enterprise stacks that say they can produce REST interfaces, but in reality most of them can't and the key problems of interface publication, versioning and testing remain unsolved.

Am I being harsh on REST? I don't think so. Its had more than enough time, and hype, to address the real problems of enterprise computing and step out of the niche, and in revenue terms it is a niche, of Web content aggregation. REST is great for that niche, but that wasn't the pitch made, the pitch made was that REST was great, WS-* sucked and REST would solve all the problems that existed with WS-*. This hype, and the smart people who followed it, has led to a stagnation in enterprise integration that is really hitting enterprise computing in real revenue terms. Its slowing the adoption of cloud computing and generally meaning that IT departments are less credible and less successful than they should be.

So one year on and REST continues to flatter to deceive.

What about Java? Well here the story is actually worse. I honestly believe that the REST crowd are a smart bunch of bunnies and are trying to solve problems, just not realising that most developers aren't as smart as them and that interface dynamism is actually a bad thing. With Java that is sadly not the case. With Mark Rheinhold now actively subverting the JCP and the debacle in the JCP generally over the last year its hard to think that the situation is doing anything other than getting worse.

What does this mean for Enterprise IT? Well a couple of things, it means that SAP, IBM and Oracle the three headed beasts of most IT estates don't have a clear future integration improvement roadmap, its all still based around WS-* as it was 4 to 5 years ago and they are all adding proprietary "tweaks" which "help" people when using their platforms. It also means that their core platform, Java, is stagnating and not getting some of the fundamental changes it needs to address the cloud and dynamic scalability. In particular the continuing "kitchen sink" approach of the Java dictatorship means that reduced profile VMs that are tuned to specific tasks (like Mobile potentially) just aren't being addressed which is leading directly to a fragmentation in the core operating platform.

What does then mean? Well with REST it means that enterprise IT shops are relying on WS-* for integration but still coming across things like vendors not having WS-Security support and with the decline of WS-I the number of "strange" defects is liable to be on the rise. Efforts around more formal contract approaches are dead. This means that the "spaghetti" of enterprise IT, which looked to improving in the first 6 years of the new millenium is actually now getting worse again. REST aimed at WS-* squarely and surely and has certainly hit it with a wounding shot, unfortunately it turns out that REST is incapable of being the replacement for WS-* it wanted to be. REST is Brutus, WS-* is Caesar.

And with Java? It means we are seeing language fragmentation and platform fragmentation which means that the support costs of IT estates are going to rise and so the on going challenge of reducing operating costs to enable investment in new solutions is swinging back against the business and towards entrenched IT estates.

I really can't see how a large scale enterprise IT programme is better off in 2011 than it was in 2005.

Technorati Tags: SOA, Service Architecture