Back in July, SnapLogic was invited to participate in an event at Oracle reviewing REST technologies and their application to enterprise middleware. Peter Laird, a Managing Architect of WebCenter organized the event in an effort to promote RESTful ideas and techniques within Oracle. It was quite well attended with over 50 attendees with about half present in the room and the others dialing in. Other participants included Subbu Allamaraju from Yahoo as well as several Oracle/BEA folks covering a wide range of topics including RESTlet, JSR 311, WADL as well as the Oracle Fusion Middleware stack.
Subbu gave an overview of REST technology beginning with the guiding principles from Roy Feilding’s original thesis. Peter and other BEA folks talked about what they were working on that could exploit REST principles
Mike gave a presentation on RESTful data services and talked about how SnapLogic used RESTful principles to address the problem. We also put together a demo that was a mashup of a data feed that came out of WebLogic Portal and a data service we built on top of Oracle. We stitched them together using Google’s Mashup Editor. Mike’s going to post the details and the links to the demo so you’ll be able to see exactly what we did.
One thing that was abundantly clear was that enterprise middleware is amazingly complex. I’ll admit to not being an expert in traditional enterprise middleware (I’m much more of a network guy), but I do consider myself pretty knowledgeable. However, when the Oracle Product Manager went through all the different elements of the middleware stack and how each supported all the industry standards and then drilled into the tool chain that was required to stitch it all together, I started to get dizzy.
When I heard him explain how SOAP HTTP bindings provided RESTful interfaces I thought Subbu’s head was going to explode. At other times it seemed as though they were speaking a different language entirely. I didn’t understand a word they were saying.
So when I saw this post by Chris Bucchere, describing the ‘ceremonious complexity’ of Oracle Fusion Middleware, I was reassured that I wasn’t the only one that felt this way.
Just as Chris suggests in his post, the same thing occurred to me during the REST workshop: Why does it have to be so complex? Surely there is an easier way?
My visit to Oracle reminded me of the vast differences between web guys and enterprise guys. With each perspective seemingly based upon physics that apply only in their own universe: Enterprise guys live in the Oracle Universe and web guys live in the Apache Universe.
For example, could you imagine Amazon’s Elastic Block Store coming out of EMC? Matter/anti-matter annihilation assured.
There may have been historical reasons and specific requirements that drove the complexity of enterprise middleware, but as we’ve seen time and time again with everything from microprocessors to operating systems, at some point the baggage of all that complexity grinds progress and innovation to a standstill.
As far as middleware goes, it really does seem like today we are back in the early days of the web. Back then, all the rules were being broken by people that didn’t know what rules they were supposed to follow. We just built things that worked that got the job done. So it seems to me that as we move out to fully distributed and cloud-based services, there is an opportunity to abandon this complexity and establish an entirely new middleware paradigm.
That’s the opportunity I see, and I think it’s pretty big.