Web services: here tomorrow, not today

Web services, while already excellent for achieving low cost interoperability through the standards that underpin them, are at least five years away from being the anticipated self-contained, self-describing XML- and web-based ‘best of breed’ ‘anywhere, anytime’ applications that will replace existing packaged suites. We’re stuck in the standards game and will be for the foreseeable future. The concept could be great: rather than buy packaged applications, businesses will harness Web services performing the functions they require on demand, stripping out cost and the problems of multiple operating systems, platforms, protocols, reformatting, integration and so on. Examples of current Web services frameworks include Microsoft’s .Net MyServices and SAP’s mySAP technology framework. It’s not that standardisation efforts that will enable XML-based business processes to talk to one another and be executable across systems, platforms, firewalls and so on aren’t underway. They are, and have been for the last two years, driven, in the open sense, by the BPMI (Business Process Management Institution) organisation – with all the main players contributing. And the commercial de facto world – companies like IBM, Sun, Microsoft, CSC, IDS Scheer, Casewise, Popkin – is busy too with prospective specifications for everything from execution languages to visualisation for business process modelling. But without agreement, further development and then commercial mass software production, the concept remains embryonic. The recent release of hitherto competing XML execution language specifications from IBM (WSFL – Web Services Flow Language) and Microsoft (XLang), now combined with Web Services specialist BEA, as the merged and prosaically named BPEL4WS (business process execution language for Web Services) signals the current reality. Competition absolutely reigns. Sun, with its alternative covering much of BPEL4WS functionality and already lodged with the W3C standards body – WSCI (Web Services Choreography Interface), seems to be being distanced by Microsoft and IBM. Yet BEA and others are involved with that too. As ever at this stage with developing technologies, the standards effort will be dogged by vested interest, hedge-betting and procrastination. Another interesting one is licensing. Will Microsoft et al requirement payment for use, or go the open source route of the BPMI, OASIS and W3C? The jury’s out on that and on the consequences. Could be that if vendors and ultimately IT departments and businesses are prepared to pay for the privilege of using the foundation techs and specs, the developers would be convinced to throw their enormous R&D muscle at development – and it would all happen sooner. Martin Owen, consulting services manager with enterprise modelling software vendor Popkin Software, is one who expects this to be the case. His advice: IT directors in companies across the sectors should be keeping a watching brief on this and helping those involved to get it right by ensuring business and IT requirements are represented. And Mike Lucas, technology manager for software services organisation Compuware, reckons those that talk in terms of a decade need to consider not only the rate of IT development, but the compelling effect of revenue opportunities of component service functionality in vertical industries on small up and up-coming companies. “Don’t take five or 10 years for granted,” he cautions. There’s also a compelling argument for users. Being able to reconfigure aspects of your IT on the fly, such that business processes and information exchange are always tuned to changing manufacturing requirements, without the cost, pain and time of today’s EAI (enterprise application integration) projects, has its appeal. Being flexible and agile internally and in terms of linking data and business processes with supply chain partners, whoever they are, can only become more important. Doing it in an open systems, ‘plug and play’ environment has got to be a better way. Either way, from users’ perspective the great white hope is that these standards come together. From vendors’ perspective likewise – but using their chosen technology not the competition’s. And so it goes on. Again, without standards, de facto or otherwise, Web Services on the grand scale cannot come to fruition. For completeness, supporting sub-standards you’ll hear mentioned are WS-Transaction (for integrity of Web Services transactions through automatic non-repudiation measures), WS-Co-ordination (handling multiple simultaneous operations) and WS-Security, now with OASIS (the Organisation for the Advancement of Structured Information Standards). And there’s the basic XML SOAP, UDDI and WSDL (transport, discovery and language support) components providing the structure of automated XML-based communication and interaction.