Don't lose sight of the "A" in SOA

Roger Smith at InformationWeek wrote a really insightful article on "The State of SOA" that reiterated that SOA isn't dead.

Still, I think we can all agree that SOA isn't exactly rocking and rolling like it was a few years ago. There are lots of reasons for that such as inconsistent definitions, unrealistic expectations and the sheer fact that SOA is a journey, not a software installation.

The article is chock full of interesting data points based on in-depth surveys that InformationWeek Analytics conducted with business technology professionals.

Along with uncovering that SOA still does have a pulse, the research points to the fact that one of the major reasons for the halting of an SOA evaluation or project is largely due to SOA having earned the reputation of overpromising and under-delivering.

Instead of pointing fingers to how and why SOA earned that reputation, let's step back and figure out what we can learn.

What InformationWeek implies - and I tend to agree - is that for a long stretch of time the focus was on the selection and evaluation of point products with less emphasis on the "A" of SOA.

Certainly you can't execute a SOA strategy without the supporting technology. And the SOA movement gave rise to some powerful offerings that are critical to the SOA's evolution. However, without a solid foundation and understanding of what the architecture can do for the business, we're merely installing products that address specific infrastructure problems yet lose sight of the bigger picture.

While the article makes it clear that SOA projects aren't going away, perhaps the recent shakeups will prompt more dialog among business and IT leaders so that both sides are clear on what an SOA - or any sweeping IT initiative - can and can't do for the business in the long and short term.

Jeff

A SOA by any other name

Now that the SOA mourning period is pretty much over, I suppose it's time to move on to the next big thing that now appears to be cloud computing.

But what are we dgoing to do about all those projects otherwise known as SOA? After all, they got a green light from management and actually delivered value so we can't just stop working on them simply because the project name has been declared taboo.

This issue got me thinking about the conversations that are happening at a lot of companies today in terms of IT budgets and SOA initiatives.

As I sat down with Anne Thomas Manes this week, it became clear that one of the reasons that SOA didn't live up to its promise was because the IT and business teams weren't speaking the same language.

IBM's Sandy Carter's pointed out this issue in 2007 when she wrote "Selling SOA to the CEO" as a guest of SearchSOA. In fact, her first recommendation was 'don't call it SOA.' (You could also argue that SOA adoption suffered due to an identity crisis. After all, the industry couldn't even agree on a standard definition.)

The reality is that SOA is not going away. It will continue on as a critical IT and business initiative though will likely earn a new moniker or perhaps just become part of the way we work.

It doesn't take much of leap to see how important the role of the architecture is to making cloud computing work. It also doesn't take much to realize that a conversation with a business exec. about architectures, ESBs and web services is fruitless when the purpose of the meting is to articulate how technology investments will deliver value to the company.

Though I fear that cloud computing may be the next misunderstood opportunity if the business benefits are not clearly outlined and if cloudy definitions prevail.

We're actually starting to see the misconceptions happening with regarding to cloud computing as Infoworld's Eric Knorr recently pointed out in his post, "Defogging Cloud Computing."

And while many believe that SOA projects will continue and will prove to be a critical part of cloud computing, we are currently stuck in a holding pattern with regard to what we actually call that project previously known as SOA.

I'd love to hear your thoughts on renaming SOA. This time, let's try to work on this name together and agree on its definition.

-Jeff

Five Reasons Why SOA's Not Dead

Welcome to WebLayers' new blog "And SOA on and SOA on." What a great time to join the fray, especially during what feels like an interminable eulogy for our close friend SOA. Not the first time I entered a market debate. People laughed when I once said email would become as central a business tool as the phone.

But what really seems to be surfacing about this whole 'SOA is dead thing' is the rather ubiquitous belief that SOA isn't dead at all.

We're not talking 'not dead' as in Elvis being sighted at a roadside diner. We're talking more about SOA having possibly moved on to its next life form.

As Anne Manes was quick to point out, "SOA is survived by its offspring: mashups, BPM, SaaS, Cloud Computing, and all other architectural approaches that depend on 'services.'"

This is followed by a slew of industry experts like David Linthicum who posits, "Will SOAs morph into private clouds?" given its crossover from SOA. Not to mention Joe McKendrick's support of this vision as well as Joe's post back in 2007 that questioned whether SOA is effectively Software as a Service. What experienced industry veteran is not an advocate of SAAS as taxonomy these days.

What I'm seeing up close is how SOA is evolving to support cloud computing and Software as a Service -- which makes me realize just how undead SOA is. Here are five more reasons why:

1. SOA companies are still getting funding. 'Nuf said from WebLayers on that.

2. The major players including IBM, HP and Oracle continue to invest resources and deliver products dedicated to SOA. In fact, IBM's latest earnings report that revenues from its middleware products – the heart of SOA - were 5.2 billion, up 4 percent versus the fourth quarter of 2007.

3. SOA projects are still underway; they just haven't been renamed yet.

4. Gartner is due to put out is Magic Quadrant for SOA governance any day now and has recently released two SOA related reports: Application Infrastructure for New Systematic SOA Application Projects and Application Infrastructure for SOA Composite Application Projects that indicate that SOA continues to be a viable initiative in many organizations.

5. It's still a busy beat assignment for the folks over at InformationWeek, eWeek, and Infoworld. Not to mention the comprehensive focus to SOA given by SearchSOA.

And I'm sure there's lots more signs of SOA's life. If you have additional evidence that SOA has not yet left the building, please drop me a line. Let's remember in an economic climate that demands doing more with less, more automation, and serious ROI's...

...the debate can't center around vocabulary it's got to revolve around results.


Jeff