Service-oriented architecture approaches maturity

The IT industry, it has often been noted, is as faddish as the world of fashion. A concept will periodically emerge that becomes – often overnight – the supposed solution to all of life’s ills, only to be dropped just as quickly in favour of a new buzz topic.

But just as the hype does not prove an idea’s worth, nor does the eventual backlash prove its inutility. In 2009, just as it became clear that a once-popular concept – namely service-oriented architecture (SOA) – had lost its appeal in the eyes of the industry’s ‘thought leaders’ in favour of cloud computing, finally some solid evidence of its value began to emerge.

The year began with an obituary for SOA, written by Burton analyst Anne Thomas Manes, who claimed that the software architecture – whereby application functionality is split into separate web services to improve flexibility and reusability – had been wiped out by the global recession. Other experts soon joined in, declaring that the extra overheads involved in developing software according to SOA principles made it a no-go in cash-strapped times.

Others countered this argument with the observation that standards-based SOA has been around for the best part of a decade and the baby was not likely to be thrown out with the bathwater, however cash strapped they were. A report published by Gartner in April epitomised this line of thinking, pointing out that, rather than being an end in itself, SOA was becoming more of an enabling technology for business process management and software-as-a-service (SaaS), allowing enterprises to accelerate cost-saving initiatives.

There is a perception, much debated, that SOA is a ‘behind-the-scenes’ technology that addresses issues that only the IT department understands. It is surprising therefore that in the Effective IT survey, adoption of SOA was as high as 13th (out of 31) in the rankings of strategies that respondents had adopted in the past year, ahead of SaaS adoption.

This finding chimes with the experience of Gartner vice president Massimo Pezzini. “In the course of 2009 we’ve seen the same level of interest as in previous years,” he says. “In addition, there have been a lot of initiatives in new verticals beyond financial services and telecoms, especially in central government and defence.” There might not have been the same buzz around SOA during the year, Pezzini concedes, but that is because “most people know what it is, how to use it and what the technology enablers are”.

Pezzini adds that SOA adoption is now beginning to facilitate further technologies such as business process management (BPM) and what Gartner calls ‘situation awareness’ or event processing. SOA is also partly behind the resurgent interest in integration and the updating of legacy systems, with many choosing to put a service wrapper round elements of functionality or add a new graphical user interface (GUI) or web front-end rather than replace older systems. Pezzini says this type of incremental initiative accounts for 75% of SOA projects. “CIOs love SOA for this because it allows them to build something new without getting rid of what they’ve got.”

Forrester Research, meanwhile, found in a major survey of 2,200 enterprises in North America and Europe that modernising key legacy applications was the top software initiative, alongside integration.

Layers of cloud

The biggest cost-saving hope, however, comes in the form of cloud computing, which almost everybody agrees is going to be a boom market for years to come.

Pezzini notes that Gartner defines three layers of cloud – at the top, software-as-a-service is “well understood, people like it”, while at the bottom layer, infrastructure-as-a-service is at least a simple value proposition, despite the security, management and legal concerns. But the big developments for 2010, Pezzini says, will come at the middleware, BPM and application server level, which Gartner terms platform-as-a-service, and Pezzini expects Oracle and IBM to join Google and Microsoft in this space during 2010.

Controversy surrounded the publishing of a McKinsey report which concluded that once you factor in all the costs, building a large enterprise application ‘in the cloud’ could cost more than doing so in-house. Others disagreed, saying that McKinsey was double-counting some costs: in one report, Aberdeen Group concluded that enterprises can save up to 20% in administrative costs.

But the real appeal, Pezzini believes, of building applications in the cloud lies beyond the cost component. “I think the value of cloud is in a whole variety of yet-to-be-discovered killer collaborative apps achieving much tighter integration between business partners,” he says.

The future of Java

While much of the controversy surrounding Oracle’s proposed acquisition of Sun Microsystems (still unapproved at the time of going to press) centred on the software giant’s possible ownership of the open source database company MySQL (see page 42), the transference of the Java development platform is arguably as significant a move in the world of software development and integration.

No-one is questioning Oracle’s commitment to Java, given the investments it has made in the acquisition of BEA and its Fusion middleware business. But there is significant unease about Sun’s dominance of the Java Community Process (JCP), including a recent call by SAP for its big applications rival to make the body an entirely open source foundation along the lines of Eclipse.

The bickering will continue in 2010, predicts Gartner’s Massimo Pezzini, but he doubts that there are genuine concerns about Java. “I think Oracle perfectly understands that the major reason Java has been so successful over the past 15 years is not its technology excellence but its openness,” he says. “We can expect that Oracle will be much better than Sun in monetising Java, but I think its biggest challenge is to communicate its sincere will to keep Java as an open standard and to win the loyalty of rival vendors.”

Pete Swabey

Pete Swabey

Pete was Editor of Information Age and head of technology research for Vitesse Media plc from 2005 to 2013, before moving on to be Senior Editor and then Editorial Director at The Economist Intelligence...

Related Topics