Product Value Management

Service Oriented Architecture has reached a critical point in its evolution. It has been extensively referenced by application vendors as their newly adopted paradigm for development, as well as by large computing infrastructure powerhouses as their platform for sustainable growth by merging acquired technologies on an accelerated basis. At the same time, the original vision of SOA, to enable fast transformation of business processes seems to be lost in between the two dominant approaches.

While SOA provides a solid set of guidelines for application interfaces development, using it simply for such purposes offers little more above and beyond of the current capabilities for application integration. SOA applied only to application integration (using standard interfaces managed via ESB message routing and transformation solutions) is nothing more then “EAI on steroids”. Many required SOA capabilities, such as service discovery, dynamic orchestrations and computing resources allocation remain unattended in this approach. The world of SOA based EAI does not look much different than before - it simply does not equates to elevated agility and flexibility of global business operations.

As far as using SOA as a information model normalization concept for merging acquired technologies, the results just seems to be somewhat more promising. Ontology based registries provide for much cleaner semantic integration of complementary business logic. However, the results of this approach remain under control of the vendor defining the roadmap, and are not necessarily exposed to the end users through widely open and accessible SOA registries. In other words, it is the vendor that decides what and when will be released in the model based environment on their still proprietary modeling platforms.

Unless, the third option, squeezed in between the two more dominant ones described above, re-emerges in a short order, we have probably seen whatever there was supposed to be of SOA. Yet, the work of many forward looking architects and developers continues under shadows of marketing clouds of gargantuan proportions casted by computing infrastructure giants who prefer the status quo. Perhaps for the sake of clarity we should stop using the acronym SOA, and introduce something much more specific - a methodology that has at least three words: business, transformation and agility front and center.

I have recently visited a project team fully submerged in designing a new (should I say - service oriented) business process information management infrastructure for a major banking institution. They have developed a three stage model based methodology whereby the team can prototype, test and evaluate and then release advanced processes for managing established and sometimes completely new offerings at the financial markets. I did not find their focus to be on SOA, or whatever we may call it going forward. The team, consisting of mostly young engineers led by few seasoned architects, follows most of SOA concepts, yet they have only fully adopted a set of principles and design techniques that enables their stated business objectives. Thus, for example, they evaluate services consumption using advanced models based on pattern recognition for forecasting demand on the computing infrastructure. Their goal is to enable end users trying to bundle a series of financial solutions to obtain a super fast quote from multiple rule based quotation engines, intelligently exchanging user profile data and marketing parameters stored in over a dozen of underlying applications. Similar approach they developed for fraud and money laundering solution, dynamically exchanging events between multiple transaction systems, intervening on behalf of the security officers by composing on a fly the sequence of events and alerting next system in the recognized pattern to disable certain services. Suffice to say their work is saving the bank tens of millions of dollars, and it is a completely sustainable model based solution that they have developed, with almost no latency in changing and adjusting to the new emerging requirements.

Yet, the team has fully deployed concepts such as standard registries and repositories, agent based frameworks, services discovery and assembly, dynamic services protocol binding, standard security profiling, and other typical SOA concepts. But, what made them successful the most, was their ability to define and deploy the entire business process within the same ontology and semantic framework fully tying model time and run time environment for dynamic promotion of process changes. I was very proud to realize that the team used some of the concepts I have published in the FERA body of work.

Back to what we think SOA is today. I met a leading architect of the large consulting and infrastructure behemoth, and discussed with him some of the work I have encountered recently, including the financial company menioned above. While he completely agreed with me that what I have seen there is some sort of best practice to utilize SOA, he also shared with me many other situations where similar solutions are emerging. The entire movement according to these new developments is very different then what we have observed with SOA so far. There is little regard to endless rhetorics of the past, perpetuated by a myriad of “guerrilla marketing” forums and their zealous pundits. However, for most of large IT vendors vying for larger control of the market, these news breaking events are not a welcome trend. They simply do not want a ground swell of independent solutions addressing what they would rather see as few proprietary platforms going forward. That way at least few of them have a chance of emerging at the top of a lucrative long term royalty base. Sharing it in two or three does not hurt them either, as it never did before.

I don’t envision the future of SOA the way that today’s leading vendors are crafting it. I also don’t see it as a free for all, open standard nirvana that has been perpetuated by well wishers and benevolent supporters without much direct interest in global businesses’ transformations. I see, however, a renewed focus on business process agility, harmonization and standardization of global value chain processes and computing infrastructure. And that may result in SOA principles simply exploding out of their current “controlled temperature egg shell” and morphing into something much more applicable and implementable - an end-to-end model based business process transformation framework. It is unclear how will the leading vendors regroup into that impeding movement. As far as I am concerned, it does not matter. Those who wait for them at the SOA cross-roads may as well be waiting forever, or accept to be taken into another direction altogether. It is a renewed competitiveness and new global business possibilities that will determine the future of enterprise computing. IT mastodons can cull SOA, or more precisely the marketing shell around it, since they have created it and funded by themselves. But, they cannot stop whatever is emerging out of its cracking shell - a yet to be named mutant of the new enterprise computing paradigm. The SOA is dead - long live the SOA.