SOA Reflects One of the Biggest Evolution of IT So far and IT Doesn’t Matter

June 23, 2005 – 5:39 pm by coachwei | Category software business model |

Service Oriented Architecture is clearly a buzz word. However, what fascinated me is what SOA reveals in terms of the evolution of the Information Technology industry,

Starting from the very beginning up to the last few years, the IT industry has been totally “vertically” focused from an architecture point of view. We always look at things within the scope of “an application”: starting from the bottom to the top in an architecture stack - data, business logic (security, work flow, business process) and all the way to user interface. The focus was to figure out how to build and run individual applications. The horizontal aspect of IT, how different application would integrate with each other, is always an afterthought.

In the last five years or so, the industry started from look beyond individual applications, moving into a “horizontal” view of IT. Now we started to look at things not from an individual application point of view, but from a horizontal point of view first. First, make your architecture service oriented – why? SOA enables “horizontal” conversations. This is such a fundamental shift – It changes the very way upon which the IT industry operated in the last 60 years.

The evolution of Service Architecture has been only a few years. Clearly, we are still in the very beginning of this evolution.

Obviously such a shift would significantly increase IT efficiency and flexibility for customers. On the other side, think the opportunities such a shift can create for vendors. It is no surprise that IBM, BEA, MSFT, SAP, Oracle etc are all racing for this. Such a shift would obviously require tools to build and maintain a service infrastructure (web services runtime, UDDI registry, web services management tools, web services security, web services messaging, development tools etc), which is what IBM/BEA/MSFT/SAP etc have been focusing on so far. On top of building and running a service infrastructure, such a shift naturally implies that you have to re-think how applications are built, deployed and managed. It means that a brand new set of tools and runtime infrastructure software are needed for applications – despite the entire industry invested 60 years and trillions of dollars in building an enormous amount of application development tools and runtime environments (If you pay attention to BEA, SAP etc, notice how they would say “composing” application vs. “building” an application).

We need to applaud to ourselves that our industry has made clear progress: we spent the last 60 years solving application level problems – now we are one level above it, trying to solve system-level problems. The scope and complexity of the problem facing IT is suddenly increased by an order of magnitude, which is why we need faster CPU, bigger storage, more bandwidth, and … new software.

We are only seeing a tip of the iceberg of what is ahead of us. “IT is dead” is plain non-sense.