Since the start of the slow recovery from the dot.com crash of 2001, the assumption has been that new investments in IT would only be forthcoming if the resulting systems could demonstrate real additional value to the business. Underlying this is the need for better alignment of IT with business requirements, and a changing business and political landscape that inevitably means that systems must become more adaptable and interoperable. Service oriented architecture represents the architectural approach to making this adaptability a sustainable reality.
SOA gained immediate acceptance in a small number of large enterprises, particularly in the financial services, government, and (in Europe) telecommunications sectors. Common factors in all of these are the complex infrastructure (often with deeply embedded legacy systems that would be a nightmare to replace), and the need to deliver rapidly new strategic, innovative services to an increasingly-demanding consumer base. The value of successful deployments (and the cost of being late to market) gave a huge impetus to SOA, and has resulted in many IT vendors committing vast sums to the development of SOA infrastructure and management technology, and for application vendors to commence multiyear projects to re-architect their application suites to be more SOA-friendly.
Now it seems that, in Europe at least, the number of SOA projects has reached a plateau. Those organisations that last year were investigating SOA have now mostly started their deployment, and those further down the road have several projects under their belts, and are accumulating valuable experience in the real-world practicalities. However, the number of organisations actively investigating SOA and starting pilot projects appears to be dwindling, rather than having SOA racing toward mainstream adoption.
Should this be regarded as a temporary blip, or does it mark the decline of this particular bubble? And what are the factors behind it?
Several important technologies went through an apparent lull in adoption: Java, Unix, and Linux to name but a few. In part, this seems to be due to a 'watch and wait' mentality by the majority of organisations that want to see evidence of the effectiveness of any new technology before rushing into deployment. At least some of this attitude is caused by expensively-gained experience from costly mistakes in committing to technologies that turned into dead-ends, or over-expensive deployments due to the immaturity of the supporting technology.
It is certainly true that in the early hype around SOA the complexity and cost was little understood and poorly communicated, but there is now considerable practical experience and supporting methodologies, in addition to a more mature software infrastructure market. It is also true that the more complex and heterogeneous the IT environment is, the more there is to gain from a service-based architecture, favoring early adoption by large enterprises.
The large number of competing vendors is starting to have a commoditisation effect on the market, bringing the entry-level price of SOA within the reach of a larger number of organisations. A number of open source distributions will make further inroads into pricing models over the next few years, meaning that most organisations that have an IT capability that includes some level of internal development will be able to cost-justify SOA more easily. At the same time, success stories from early adopters are starting to filter through, providing the incentive to take a further look.
However, these factors may not be sufficient to bring a significant acceleration in the adoption of SOA. The missing element is business sponsorship, and this means that better communication of the aims and benefits of SOA needs to be put in place. Treating SOA as an issue for IT alone is never going to achieve this. As a tool for IT, SOA merely gives a more cost-effective way of managing integration. As a capability for the business, it is at the heart of business/IT alignment with both short- and long-term deliverable benefits including the optimisation of business processes, the ease of integrating new business partners, the ability to create rich composite applications and mashups very quickly, and the ability to be more selective in the outsourcing strategy. Agility in each of these is the key to being able to gain a first-mover advantage as the business environment continues to evolve.
Most IT organisations have yet to address this need to involve business management more closely in the development of their IT strategy, and until this happens SOA is likely to remain a minority architecture, benefiting just those that take the time to understand that this really is the meeting point between the IT and business worlds.
Butler Group is a ComputerWire sister company and part of the Datamonitor Group.