SOA Strategies with Neil Macehiter

Posted March 14th, 2008 by Joe Pendry

macehiter_sm UK analyst Neil Macehiter of Macehiter Ward-Dutton took some time to discuss his views on SOA, IT-business alignment and web services.

Neil’s areas of expertise include enterprise architecture, service-oriented architecture (SOA), web services, virtualisation and identity management. He has acted as an advisor to leading vendors, including IBM, Oracle, Microsoft and Sun Microsystems; and to large IT user organisations in Government and the private sector.

StackSafe: How did you become an analyst focused on enterprise architecture and SOA?

Neil Macehiter: This is actually my 23rd year in the IT industry! I became an analyst back in 2001 when I joined Ovum, having spent most of my career working for a variety of large vendors (Oracle, Sybase, Sun) and start-ups (Autonomy, Zeus) in consulting, pre-sales and strategy roles.

At Ovum I was a Research Director running a team of analysts looking at software infrastructure – application servers, EAI, application development – and soon got embroiled in the world of web services and WS-*. I guess that was when I first dipped my toe in the SOA water.

When we founded Macehiter Ward-Dutton in February 2005, the web services story had moved on and everyone was pushing SOA but the theme of the story remained much the same: software development and EAI on standards.

Here at MWD, we focus on IT-business alignment and we quickly realised that service orientation is a key enabler and since then we have been focusing on SOA from the broader perspective of a service-oriented approach to investment in and delivery of IT capabilities in line with business outcomes.

When I was at Ovum I worked with a number of large public sector organisations both in the UK and Asia Pacific. Many of these organisations were looking at EA methodologies like the Zachman Framework and TOGAF so that was my introduction. Since then we have been thinking about the role of EA in IT-business alignment; how it relates – or should relate – to SOA. Another important driver has been the research we did for our book “The Technology Garden: Cultivating Sustainable IT-Business Alignment”, which covers some important aspects of successful EA.

StackSafe: What do you see as the biggest trend or advancement in SOA during the past five years?

Neil Macehiter: Wow. SOA has moved evolved so rapidly in the past five years so it’s pretty difficult to call out one particular trend. That being said I would call out the growing recognition of the importance of SOA governance and a maturation in the way the industry thinks about SOA governance.

Governance is about doing the right thing as well as doing things right. In the early days of SOA and web services much of the focus was on design time governance but if you think about services in the real world e.g. your cellphone service it’s actually something you experience. SOA governance should be about the end-to-end service experience and so needs to be about the operational phase of the lifecycle too.

This in turn means that organisations are going to have a very clear understanding of the operational, non-functional aspects of their services as well as the functional aspects e.g. the “how well” and “how much” as well as the “what”. Services contracts should reflect those aspects and become policies at runtime.

StackSafe: Will Web 2.0 standards make SOA more difficult to manage and why?

Neil Macehiter: I am not totally sure what you mean by Web 2.0 standards. I think that SOA and Web2.0 are different sides of the same coin. Web 2.0 is best characterized as the next phase in the evolution of the web – web as library; web as ecommerce and now “web as a place”.

Web 2.0 is about organisations exposing and exploiting capabilities across traditional boundaries, be that to enable collaboration and dialogue with customers and partners; to exploit hosted capabilities and SaaS; be that to perform B2B transactions – essentially this is changing companies – they are becoming uncompanies.

SOA is about organisations exposing and exploiting existing capabilities behind the firewall. Both are predicated on the use of open interfaces, standards, data formats etc but Web 2.0 tends to be less rigorous and structured – what we refer to as “integration through the back door”, whilst SOA is “integration through the front door”.

I am not so sure it makes sense to say that Web 2.0 makes SOA more difficult or SOA makes Web 2.0 easier – I think they will blur at organizational boundaries and meet in the lounge, somewhere between the back and front doors.

StackSafe: What is your opinion of the alignment between developers and IT operations with regards to SOA?

Neil Macehiter: Historically this has been poor. As I mentioned about the alignment is beginning to improve with the maturation in the thinking about SOA governance. Vendors such as HP and IBM are beginning to recognise the need to close the gap today. Microsoft, too, has acknowledge it as part of its Oslo strategy.

There’s still a way to go though. A service-oriented approach to IT requires IT operations to abstract the servers, storage, networks and focus on the services those underlying infrastructure capabilities support and enable. That provides a common language to facilitate discussions with their developer counterparts and will provide a common way of thinking about things that extends from development through to operations.

Service contracts also have an important role to play here since they define the functional characteristics that developers must provide and the non-functional characteristics that IT operations must enforce. They need to be shared and accessible throughout the service lifecycle in some form of shared repository.

StackSafe: What problems do you hear most frequently (either SOA-related or not) from the companies you advise?

Neil Macehiter: The challenge of ensuring that the way they invest in and deliver IT is aligned with business outcomes. SOA certainly has a role to play but organisations are struggling to see how that fits with enterprise architecture initiatives, how to articulate it in terms that the business understands and in a way that the business can see the value of. We’re also seeing a lot of confusion around the relationship between SOA and BPM. BPEL is not the answer!

StackSafe: Which vendors do you think have the most mature viewpoints on SOA?

Neil Macehiter: That’s a really difficult question when you consider that our perspective on SOA is about the service experience throughout the service lifecycle, from initial design through to operations and beyond. So many different types of technology and associated vendors have a role to play, from application lifecycle management, through ESBs, web services management, identity and entitlement management, registry/repository etc etc. It’s about much more than software development and integration where many of the vendors who took the early high ground in SOA came from.

I am not dodging the question. Rather I think organisations need to think first about what they are looking for from their SOA initiative – is it integration? Is it delivering new, user-centric capabilities? Is it exploitation of SaaS etc – and then to think about the implications of a service-oriented on the way they do things today. That will help them to identify what they need and therefore which vendors viewpoints and, equally importantly, vision fits with what they need today and where they are going.

This is something that’s brought out in our free SOA strategy planning tool which considers multiple technical and non-technical aspects of an SOA initiative.

StackSafe: Are there significant differences in approaches to SOA between European based companies (specifically UK) and other areas?

Neil Macehiter: In our experience, architects (and I don’t mean developers on the next rung of the career ladder) play a more significant role in SOA initiatives in Europe.

StackSafe: If a company could focus on one area to improve SOA, where would you recommend they begin?

Neil Macehiter: I would recommend they begin by thinking about the end-to-end service lifecyle and its implications for the way they do things today. (They could do worse than give our SOA planning tool a whirl too. )