Speaking the language of business with SOA

Walking through terminal 2 at Heathrow on Monday morning, I saw one of the What Makes You Special* adverts near my gate.

I’ve been reading Sandy Carter’s recently-published book on Service Oriented Architecture (SOA), The New Language of Business: SOA & Web 2.0. I’m about halfway through, so it’s possible that my opinions will change before I reach the end… but I thought I’d make a few points about it anyway.

The book is divided into three parts. The first talks about the business context – driving home the key point that SOA is all about business as well as IT. The second section talks about flexible IT to enable a flexible business, and has chapters covering the key concepts of SOA, governance, service granularity, and how Web 2.0 relates. The final part of the book (that I’ve not fully read!) is about how the concepts can be applied, and uses IBM as a case study as well as presenting some key “don’ts” for SOA implementations.

I have to say that it has struck a lot of chords with me. It probably all started with the fact that the first case study (on the importance of business processes) used a banking payment processing example, and I’ve done a significant amount of work in that area in the past five years! I’m also impressed by the business focus. As a techie I need to consciously switch my mind away from “hey, look at this cool new product feature!” towards “how can we use this stack of technology to improve business effectiveness”, and Sandy makes some great points in the early chapters that helped me to to do that. In particular, chapter 3 covers Component Business Modeling, which is an important concept in helping to align business and IT.

Quick aside: driving home from Sheffield last week, I listened to the recent Redmonk Radio episode about “Incremental SOA with NetManage”. I’ve been consulting on Enterprise Service Buses (ESBs) – yes, pretty exclusive in the form of WebSphere MQ and Message Broker, but the specific form shouldn’t matter too much – for several years now, and a lot of what the guys talked about on the podcast made sense to me. It is common sense stuff. Although SOA is a potentially revolutionary model, trying to implement something as a “big bang” is generally not going to work – pick a smaller chunk of the business and technology, implement a small project (possibly one that does some form of wrapping of some legacy code to enable it as a service, as Archie Roboostoff of NetManage mentioned on the podcast – that can be a good demonstration of the value of the middleware), and build up from there.

Why mention that podcast? Well, as I listened to it, I was thinking “surely this incremental thing isn’t news?”, and I also decided in my own mind that SOA could be revolutionary, but it needed to be evolutionary too[1].

Sure enough, Sandy’s book talks a lot about choosing a place to start and building out from there. In the introduction to the book, IBM’s Steve Mills states that

An SOA is an evolutionary approach to building IT systems that is focused on solving business problems.

In a later chapter Sandy goes on to state that SOA is an evolution not a revolution (which made me jump since I’d essentially had the same thought last week – but again, this is common sense stuff to me). She also writes

A key to any SOA entry point is to start with a discrete project and then progress over time.

There are some pretty handy checklists in the book, too. Picking just a few of the “10 secrets for success” in implementing SOA, we have some of my favourites: #1 Get executive sponsorship, #2 align the troops (i.e. get people educated as to why the business is going down this road), and #7 … hop on the ESB!

Tying all of this into the Web 2.0 world, one example of flexible business that I picked out from the early chapters was

… the ability to … provide consistent multichannel access for customers to increase customer loyalty …

A couple of years ago, this was all about portals. Today the portals are still there as the user interface to the SOA, but enhanced with feeds, mashups, and other technologies which enable the access and improve the user experience.

If you are new to the concepts of SOA and wondering what the fuss is about, this is a useful book. It’s also a good summary / distillation of some of the key issues that we’ve come across over the past few years. Overall, worth a look.

Now if you’ll excuse me, I’m off to reconfigure an ESB for someone 🙂

Related link: Books by Sandy.

Disclosure: I received my copy of this book at no cost, but was going to buy a copy anyway.

[1] an unfortunate by-product of this is the way in which many of the standards are layered on top of one another, or reinventions of previous concepts – as an industry we seem never to want to throw anything away, so it gets wrapped, and then something newer comes along, so we wrap that last thing… still, at least XML has made this both easier, and more verbose 🙂

Technorati tags: Service oriented architecture, SOA, ESB, Enterprise Service Bus, Redmonk, IBM, Netmanage, Sandy Carter, Steve Mills, Web 2.0