There is an unfamiliar tone in the voices of IT directors as they describe their early experiences with web services. It sounds like pride. Maybe even cockiness.
Could these be the same executives, who bought into – and were burnt by – oversold technologies (CRM, ERP, e-procurement and others) that earmarked IT as inflexible, slow to deliver, and costly?
Their elevated esteem stems from what they have already managed to achieve with web services, the revolutionary standards-based approach to applications that packages functions as discrete entities and enables these to be ‘discovered', freely bound together and executed over the network. The industry hype
surrounding the technology is all too familiar, but as organisations start to talk about their early projects maybe, for once, the technology is delivering to expectations.
From banks to government agencies, the anecdotal evidence is there: that developments that would have taken months with conventional technologies are going live in hours; that projects are coming in under budget; and that upfront costs are negligible.
Could this possibly be true? Ask JP Rangaswami, Global CIO at investment bank Dresdner Kleinwort Wasserstein (DrKW): "Web services is by far and away the only way to go. The kind of bill CIOs have had to pay in the past for highly visible, expensive, gloriously celebrated failures is now much less likely to happen."
Or John Picanso, CIO at the Colorado Department of Agriculture: "This is a true case where the technology is going to deliver over the hype." No-one is saying that web services is a panacea for all IT's weaknesses. And indeed, many of its early applications are relatively modest in scope. But where are organisations seeing strong results, and where are they finding the technology lacking? What costs are involved? And where are the inevitable pitfalls?
The concept of web services is so broad that it hits to the heart of the shortcomings of the technology industry – shortcomings that have remained largely unaddressed for decades.
"What does the [business] really want?" Rangaswami recently asked at a CIO summit run by The Economist. "It expects its [technology] acquisition and replacement cycles to extend much further, rather than having to keep writing off technology every five to seven years. It expects to have middleware layers that simplify application integration and ensure speedy time to market. It expects to have visibility into the cost of ownership. But more than anything, it wants to be able to retain and acquire new customers by providing them with [alluring propositions] such as personalisation, ubiquity of access and so on."
"In every one of these cases," continues Rangaswami, "I can happily say that web services allows you to reuse software components and integrate like no previous technology; that the time to market is reduced and the ability to customise for the end user is greater than ever before; that total cost of ownership is reduced, because you are not building a large ‘configure-your-own-silo' project and it is likely that the cost associated with maintaining applications drops; and that projects tend to be much smaller and iterative."
He can say that with some authority as DrKW was one of the very earliest users of such technology, in the form of Microsoft's XML Web Services. Among other projects, DrKW has built a foreign exchange (FX) platform for its investment banking customers using web services. First, it broken down the FX value chain to identify the core business process, it then adapted the various existing applications that supported each of these to ‘expose' their functionality as a web service. Having exposed this core functionality, it was able to bind together a customisable application that integrated FX functions from various systems.
The project resource commitment: one person, 40 man-hours. "Had we tried to do this [with conventional programming approaches and technologies] it would have taken at least a month," says Jan Jones, VP of IT at DrKW. "And we [then] wouldn't have been able to customise it for the next client so easily."
The underpinnings for such flexibility is that, for once, in an industry that has often seen standards as blunting competitive edge, the standards are actually in place before the technology is applied. Furthermore, the key standards are universally accepted by the industry's giants including Microsoft, IBM, Hewlett-Packard, Sun, Oracle, SAP and Siebel.
Three standards in particular – all based on XML, the extensible mark-up language – make web services a reality: SOAP (Simple Object Access Protocol), for transmitting XML-encoded data and remotely accessing services/objects in a platform-independent way; UDDI (Universal Description and Discovery Integration) for registering and discovering services; and WSDL (Web Services Description Language) to provide an XML grammar for describing available web services.
Reuse, reuse, reuse
That standards foundation has put one word at the top of the agenda: Reuse.
"For us web services is all about the reusability of commonly used business functions," says Tesco.com's head of ecommerce development, Jon Higgins.
When Tesco.com decided to redevelop its online shopping site to support PDAs, digital TV and additional front-end clients other than PCs, it found that by wrapping its existing business objects in SOAP, it could expose them as web services to any kind of different client without any changes to the back-end logic. The upshot: a few hours work rather than months (see box, In practice: Tesco.com).
"From an operational perspective, [it] makes strong financial sense," says Higgins. "We can develop an application once and then reuse it across different channels without having to write application [logic] for each platform. Web services is enabling [us] to get to market quickly and obtain an immediate return on investment."
Speed of implementation was even more of a consideration at the Colorado Department of Agriculture (CDA). Towards the end of 2001, the department needed to find a better way of distributing information gathering on a distressing problem: Chronic Wasting Disease, a transmissible spongiform encephalopathy akin to BSE, that is reaching epidemic levels among the state's herds of elk, with as many as one in 100 infected.
The business case was to come up with an application that would take much of the disparate information gathered by the government department, other agencies and academics as part of their monitoring of CWD and disseminating this for quicker decisions on controlling the problem.
Calling on the design and implementation skills of professional services and software company Compuware and taking Microsoft's Visual Studio .Net and the .Net Framework, the CDA built an application that could expose parts of the agency's data platform (an existing system, Enterprise Integration System) as browser-based web services.
The web service application went live in March 2002, just a few months after its inception, at a total cost of around $100,000 for all systems, software and services. "What it gave us was the ability for everybody to see that data. If we had gone with the true client/server approach, the project would have taken probably six to eight months," says the CDA's CIO John Picanso.
The European operations of PC maker Acer had a similar problem, but some very different data to disseminate. The company has 12 standalone enterprise resource management (ERP) systems handling order processing, inventory control and financial management systems in 12 different countries. The business problem: Management and fulfilment teams in Holland needed consolidated reports and analysis that provided a pan-European view of the business.
After rejecting enterprise application software, Acer decided to create its own application using web services software from Silverstream, building web services components that accessed various parts of the ERP system and extracted information.
An example of how successful the $1 million, eight-month project has been: the standard cost update used to take two days, now it takes two hours, says Howard Cheung, CIO of Acer Europe.
Clearly what these examples show is that web services is being most readily excepted as an integration platform – especially within organisations.
That may simply be because they are addressing the biggest headache first. Analysts estimate that 40% of the cost of major projects today is integration-related.
Even when the application extends out to an organisation's partners, integration is still the catalyst for web services adoption.
"We see web services as an integration tool and nothing further," says Simon Harris, head of development at LevelSeas. The e-marketplace for cargo shipping space has used BEA Systems' WebLogic Workshop to build a ‘contact synchroniser' web service that allows the exchange to capture and expose user desktop contact lists and thereby help find a match for available cargo space.
What is so different here from previous technology adoption is that upfront costs are often negligible. In most cases, the capabilities come with scheduled upgrades of software that is already in place.
"None of our web services work incurred any direct software costs," says Harris. And echoes Higgins of Tesco: "When .Net components became available, the facility was there to build web services simply as part of our upgrade."
That situation could have a severe impact on the software industry, as it encourages far more in-house development and the purchase of fewer packages. As it shows tangible results, a sense of power is moving back to the IT department after years in which the major technologies – ERP, CRM, and so on – were sold directly to the board, with IT merely as the implementer.
And this is only the beginning, says Tesco.com's Higgins. "It feels a bit like the birth of the Internet when the web arrived. Almost from day one, it became clear that the web was going to work because there was a well-defined series of fairly open standards. So anyone could get a web server, make it spit out http and talk to anybody's site. This is the same feeling: that with a well-defined series of standards, people can build something that is going to be ultimately powerful."