I have discussed SOA a lot over the years – I am obviously a proponent. From my previous blogs - I do believe an effort must be made to make SAAS and SOA compatible.
- SaaS must be more than “Applications as a Service” – largely what we see today. We must encourage application vendors to expose their applications as loosely coupled services. The writers of “SaaS Manifestos” (myself included) must speak to how SaaS should work in the wider context of an interconnected business. Multi-tenancy is a necessary condition for SaaS to be economically viable and well built – but it is like asking a car dealer “Is the car reliable?” when you really want to know how it handles.
o I translate this into “SaaS Imperative #1: SaaS cannot be black box – we learned our lesson the first time with on premise software”
- SOA must embrace SaaS. SaaS vendors – even where they embrace service orientation – are not enough. All the services must be orchestrated into the services and processes that the business defines as strategic.
o So here goes “SaaS Imperative #2: SOA enablers have to be ahead of the game – move to the Cloud nowl!”
But I began writing this blog to ask a completely different question – “SOA … what next?” SOA has been around a while and its implementation has been a lot about back office orchestration. But with Web 2.0, alot of the ownership of business processes are shifting from business to autonomous consumer or supplier – moving from “macro” processes of well managed companies to what appears to “micro-chaos” world of consumers and small suppliers.
SOA must expand to manage the phase transition from “macro-order” to “micro-chaos”. Some macro-companies handle this naturally – Apple (micro-applications), Amazon (micro-vendors) and Microsoft (micro-ISV’s). Most companies don’t. SOA must evolve to help the majority of companies that do not. And it must be more than Google Wave, more than Amazon payment services.
I am convinced that SOA which now concentrates on exposing services and processes to a business’s partners (including consumers) must now learn how to: assist, persist, assure, automate and incorporate customer’s micro processes.
In other words – SOA must host all parts of the business model – regardless of the owner of the “Service Bus” (if that is going to be the term). Successful businesses will look for SOA technology that allows them to adapt and adopt the chaos of the small “actors” in their business models and yet maintain order.
My guess is that it will be some form of “Collaboration” – not Facebook, MySpace or LinkedIn– where closed garden collaboration is the business – rather where collaboration is open – where any business and any consumer can paricipate.
IBM and Microsoft will take some leadership. As with the browser wars – the regulators will be suspicious as current collaboration sites fear a destiny much like NetScape.
Some companies – such as Oracle – who are so squarely rooted in the Macro and black box applications (and whose desire to further SOA have never been clear) – who knows...
Oh yeah, back to SaaS. Distributed IT has automated business for 20 years – but it is too often expensive and wasteful. Like SaaS, Ford created the centralized auto assembly line to maximize economies of scale. But Ford produced Model T’s. Many decades later – the distributed production model is bringing back the built to order automobile.
In IT SOA saved us from static automation once – we can’t wait another 10 years to make SaaS dynamic. This is especially important as business models make a true step change to address the micro.