Whatever happened to SOA?

Not too long ago Service Oriented Architecture (SOA) was the hottest IT trend around, at least if measured by the sheer amount of attention or hype that was associated with it. The fate of SOA is interesting for a number of reasons:

• It helps to illustrate the nature of Hype Cycles in IT and how they influence both industry and individual enterprises.

• It helps to provide context for why Cloud solutions have now become the "killer solution" (it can't be described as an app) that most folks are examining today.

• It is an interesting case study in regards to the difference between product-based Hype-Cycles and architecture or standards based hype cycles.

So what did happen to SOA? Strictly speaking, it's still around and going strong, however it is no longer at the top of the 'must have' or 'must do' lists in most IT departments these days. Many SOA initiatives took the bridge to nowhere or have been merged into newer or perhaps more specific enterprise goals and objectives. Before we go any further though, let's recap what SOA represents:

1. SOA is not limited to one product or product family.
2. SOA encompasses both capabilities and process.
3. SOA is very focused on a key set of (mostly W3C) interoperability standards, notably SOAP, XML, WSDL and UDDI.
4. SOA is not merely a solution but also a solution strategy and architectural approach. This last point is both critical and very subtle. 
5. SOA has an systems integration focus, not merely a software development one.

This diagram illustrates a number of the Use Cases associated with SOA Adoption

 Each of the last five points is important and several of them are relatively complex issues. Essentially, taken together they tend to imply that anyone wishing to deploy SOA in a significant way would in effect be attempting some form of enterprise-wide transformation. That's a big deal regardless of the technology or technological paradigm targeted in the transformation. Change is never easy - it's less so when dealing with an entirely new way of managing your IT portfolio. These are some of the reasons that SOA is not the driving force in industry it was just four years. So, let's look at what did happen with SOA:

• SOA was and SOA is hard. There's no silver bullet solution - all options require a lot of work and some risk.

• SOA initiatives usually become Systems Integration exercises. Many providers or organizations without that systems integrations background attempted SOA solutions and quickly realized it's a skill that takes years to develop (both personally and organizationally).

• There's more than one way to do SOA - in fact there's quite a few ways. This implies a wide array of possible tools or software options. Navigating those decisions is confusing to say the least and since this happens up front it often short circuits projects before they ever get off of the ground. There's not one piece of software associated with a proto-typical SOA solution.  

• Often times key elements of SOA infrastructure are not implemented as part of an initiative. Common examples of this include failure to deploy a SOA Registry or Governance paradigm. This tends to limit the effectiveness of the solution and eventually restricts the scope of the entire solution.

One of the reasons people were initially enthused about SOA was that it seemed to represent an easy path to bypass many of the ills associated with the ad hoc heterogeneous sprawl that had become the typical IT enterprise by the mid-2000's. That initial assessment was a misconception, but so too are similar assessments being made now about "Cloud Computing" and "Big Data." Behind the Hype and the marketing buzz, there are certain underlying problems that all IT trends are focused on resolving. More often than not however, no single trend is positioned to solve the challenges alone. We will talk more about Cloud Computing, Big Data and SOA in future posts and presentations. Also be sure to check out our Technovation Glossary page.

Copyright 2012, Semantech Inc. All rights Reserved