SOA is alive and well no matter what its called
A recent posting by Anne Thomas Manes of The Burton Group —
Has declared “SOA is Dead”, yet “long live service-orientation”. This has sparked quite a flurry in the blogosphere, on the Yahoo!Groups Service-orientated-architecture discussion group, and on SOASocial
Miko Matsumura has now labeled it “The Architecture Formerly Known as SOA” in his post “SOA is Dead! Long Live Whatever!” and he referenced this photo –
If I actually believed the controversial statements, I would prefer to call it the Architecture-until-recently-known-as-SOA –
However, based on my travel schedule over the past several months, which has been about 90% of my time going to meet with customers who are actively engaged in funded SOA projects, I would say the architecture-until-recently-known as SOA is still-alive-and-well.
But seriously folks….
Anne is a friend I have known for a long time and I continue to have a tremendous amount of respect for her, but I think she is off the mark on this one.
One could argue that there are discretionary projects being put on hold due to budgeting restrictions, but I haven’t run into that. If its happening then I doubt that has anything to do with whether they are SOA. How can you distinguish it anyway? Almost every major project in IT has some kind of SOA element to it these days.
I would think during a recession period, organizations would prefer to cut costs and the architecture until-recently-known-as SOA would be the solution to bring down application development and maintenance costs through increased reusability.
While the title of the blog is contentious – the thoughts presented towards the 2nd half of the article are not all that outrageous.
Towards the middle of the article, it starts to sing a common tune – “the requirement for service-oriented architecture is stronger than ever.” So, in effect, she believes in the promise of services, but think its time to cut through the hype and define SOA as not just a software/technology, but a mindset which requires shift at organizational level as well. Well put, and I don’t think you would find any disagreement.
So where is the dead part?
So..Ahhhh, what does it matter what we call it anyhow? OK, I’ll bite.. SOA is Dead. Long live service-orientation! ….and architectural patterns and practices to ensure that an IT organization has consistency across applications that are built using service orientation….Oh and I guess we’ll need a means for ensuring that those services and their associated artifacts can be stored, versioned, secured, and accessable from a variety of interface technlogies and protocols. And we can’t build it all from scratch so we’ll need to be able to leverage our investments in existing application assets….then we’ll need to compose services together, and perhaps even orchestrate the interactions between them….We’ll also need to have a working plan in place to communicate between IT and the business to ensure that the right services are being built at the right level of granularity so that applications, which leverage those services, can have the flexibility and agility to meet the increasing and ever changing demands from the business. And a way to govern that process….
In another retort, Jack Vaughan of SearchSOA.com responded with –
It’s good to have give-and-take and controversy. But this
has fully played out in the SOA space. It’s everyone’s job to build
good apps and good integrations. If you feel like discarding SOA as a
buzz word, go for it – no need to ‘kill’ it. The claim that “SOA is a
great failed experiment” is extraordinary hyperbole, and such
hyperbole is the enemy of getting the job done.
In all fairness to Anne, during a conversation with me on Yahoo!Groups service-orientated-architecture discussion group she clarified her message when she said this –
My point is that IT people should stop using the term “SOA” when
talking to business people. They should not try to sell architectural
concepts to business people. They should focus on selling solutions
that deliver value to the business. But architects should still apply
service-oriented principles to the solutions that they deliver.
I for one am going to continue to call it SOA until I hear a better reason not to.
Dave