Dennis Howlett has added some comments to my recent blog on Web2.0 and SOA. Dennis makes three points, one of which I agree with, one I don’t and a final one which highlights an interesting issue. To paraphrase Dennis’ points:
1. SOA is already a dead concept: “The real problem lies in the term itself, its vagueness and its associations with blue sky projects that demonstrate little more than debatable business value.”
2. SOA is poorly named and would be better called Service Based Architecture.
3. SOA projects need to be applied to specific edge projects – as Web 2.0 projects have been.
I won’t argue point 2 – the Service Oriented term echoes (of course) Object Oriented and indeed service based would be a closer match to what it is about. However, the fight is long over on this one, and we will have to live with Service Oriented.
I will argue with point 1 – Dennis’ points of evidence are application vendors such as SAP saying it is dead and EAI vendors claiming that it is nothing new. SOA is potentially deeply threatening to application vendors such as SAP who sell tightly integrated suites of applications as SOA in its essence encourages mix-and-match. On the other hand, EAI vendors have embraced SOA and following the usual software marketing approach claim that they were doing it for years before the term was coined – we simply didn’t notice.
Moving onto point 3, Dennis is absolutely right that SOA projects need to be prove value and there is a tendency to redefine the term to stand for Seriously Overly Ambitious – turning them into big bang projects which are hard to fund and hard to demonstrate RoI on in many cases. This is not necessary – SOA implementation can and should be done incrementally. Furthermore Dennis hits the nail on the head when he says that one of the distinguishing features of Web2.0 has been the strong focus on productivity: it is easier to work with and therefore encourages innovation. Why doesn’t SOA seem to follow the same path? In our keenness to stress the architecture, as proponents of SOA we risk missing the point that the tools and approaches used can and indeed must be productive if we to prove the benefit of SOA quickly enough to maintain momentum within the organization.
Subscribe to feed
Of course SOA threatens SAP, that's why NetWeaver exists. And I can see where the EAI boys are coming from. But I remain of the belief that as a concept - and I should qualify - for explaining to business users - is dead. As an architectural concept, it still has life. But only one project at a time. And that kinda defeats the object of the exercise.