SOA OD at Sapphire?

I was at Sapphire in Paris last week, and found the process pretty useful. Met a lot of interesting people, and the client I was there with are considering HR initiatives, so some of the technology (e.g. Duet) looked potentially useful for them.

But Sapphire isn't really a techno-fest like many other IT conferences. A couple of characteristics that you couldn't help but notice and that aren't really techie at all were a) the huge number of vertical industry-solution tracks covering what seemed like every business vertical there is, and b) a huge amount of Enterprise SOA (SAP ESA as was) activity and information. Both were impressive, but the latter of course ran the risk of being a bit overwhelming.

But this is really a characteristic of the software marketing machines rather than a Sapphire-specific issue. Vinnie Mirchandani at deal architect has posted that he believes SOA is being over-promoted and maybe over-engineered (my words not his). Thomas Otter's also just got back from Sappire Paris and has has blogged that he is in some agreement with Vinnie. It is apparent that there is a risk that the heavily software-centric SOA marketing obscures the facts that it takes a lot more than some new technology, or some new approaches to applications, to produce business-valuable results from SOA. This is similar in some ways to an old posting of mine on the real issues making or breaking SOA.

It seems that for many enterprises, the major value to come from SOA isn't from the 'S', but rather from the 'A' in SOA. For many, this generation is the first time they've seriously thought about architecture at all, at least above the project-level, sometimes above the tin-level. Many will benefit from the flexibility and focus of service models, but for some that may be second to more basic value-propositions, at least in the short term. Fundamentally the focus on achieving value beyond a silo, driving commonality and collaboration, no longer sub-optimising and managing technology assets on a business-lifecycle rather than a technology one may be where much of the value comes from for some.

But I'm optimistic about SOA. Even though there were a huge amount of ESA sessions at Sapphire, I was pleasantly surprised with some of the pragmatic insight emerging (rather than the all too common "SOA will make you more flexible / agile / attractive-to-the-opposite-sex"). They were also giving out copies of Dan Woods' new book on Enterprise SOA, and I was also pleasantly surprised at the next generation of IT management and operational insight they have added there since his last. Ron liked it as well. However, the true business view of how service-orientation can add value to businesses and what is necessary to achieve it (the service-oriented enterprise if you like), was more notable in its absence.

James Governor has commented on how SAP have generally had more insight and more communication skills about the business angle than the other IT vendors, and he's right. But that said, I do believe that even business-minded IT folks and IT vendors need to be careful with the dogma about services. I believe that the real services thinking needs to come from business people if you are to avoid the cart attempting to lead the horse. Of course, the perfect solution is for business people seeking to innovate and transform their business to include those business-minded IT people in the process, as joint teams are always going to have the best chances of identifying the optimal solutions.

One last bonus at Sapphire was the Chief Architect Session on Tuesday afternoon. SAP spoke about ESA and their new BPX (Business Process eXperts) community, which was good to hear about, but there was also a refreshing item where Derek Prior from AMR talked about the importance of EA in a) providing leadership, methods and skills to actually execute on SOA, and b) bridging the business and IT gap for collaboration in SOA. Anyone who's read posts of mine before will no doubt guess how much I liked this.

Of course, this EA Session is all part of the big applications vendors increasing interest in EA (as Neil Ward-Dutton has blogged before about here), but the fact they are doing so is a great move, for them and the industry. For far too long architecture was viewed by the apps-people as being separate from the applications world, and even more separate from the apps-vendors' world. It was seen to have come from an infrastructure design and bespoke software development background. But this artificial historical division just meant that each vendor-island created its own silo of narrow architect-type roles (staffed by often excellent business-focused individuals). I believe that whilst of course you will always need specialisms, these hard divisions can't be sustained as the next generation of Enterprise IT (whatever it's called) breaks down these silos and looks for bigger and additional value-add.

Technorati Tags: SAP SAPPHIRE06 SAP ESA Enterprise IT Enterprise Architecture SOA Service-Oriented

Relating your services to your capabilities in SOA

So what exactly is a Service? If we are seeking a common definition of what SOA is (and much is made of the lack of this on the web), then we could at least start by defining the 'S' (in SOA).

There is some emerging consensus in some SOA communities that 'services are a way to gain access to, or make use, of a capability'.

What does this mean? Is it just semantics? Well actually it is important, and the reasons why are related to why services are a useful concept (and implicitly, when they're not!). I've posted a couple of thoughts as to why up on Nigel Green's Services Fabric blog.