Web services

Web services may be the El Dorado of the software industry. Just as Sir Walter Raleigh and countless prospectors looked for the city of gold, vendors and corporations alike have sought out ways to enable application accessibility across the Internet from a variety of devices -- without having to worry about the infrastructure involved.

The notion of Web services exploded in June 2000 when Microsoft Corp. announced its .NET strategy for Internet-based applications and unofficially dubbed such applications "Web services." Since then IBM Corp. has touted technologies it already has as Web-services enablers and, more recently, Oracle Corp., Sun Microsystems Inc., and Hewlett-Packard Co. proclaimed that they too have been working on Web-services technologies for some time.

With all of the major software vendors now backing Web services, little doubt remains about the general direction in which the industry is heading in terms of the dominant environment for business applications. "Web services are a very profound advancement for e-business. [They] can change the Internet into a platform for application integration," says Scott Hebner, IBM's director of marketing for WebSphere.

What are Web services?

Although most of the important vendors have pledged some sort of allegiance to Web services, and all have detailed plans, none has clearly articulated exactly what Web services will be. "There are so few examples of Web services in action that it's difficult for developers to understand the value proposition," says Rick Ross, president of Cary, N.C.-based Java Lobby.

That may be true because the technology is in a state of infancy, and won't begin to mature for at least another two to three years. Early examples of Web services include on-demand delivery of stock quotes to a cell phone or pager -- but such bare-bones capabilities are merely the tip of the iceberg.

Web services are software components that represent business functionality that can be accessed by users -- via applications or another Web service -- using standard protocols. More importantly, a Web service may combine several applications that a user needs, such as the various pieces of a supply-chain architecture. For the end-user, however, the entire infrastructure will appear as a single application.

Early supporters of Web services say that almost any application imaginable can become a Web service.

"There are going to be more Web services than Web pages," says Frank Moss, co-founder and co-chairman of Lynnfield, Mass.-based Web platform and infrastructure provider Bowstreet Software Inc., and former president and chairman of Tivoli Systems Inc. "Companies will come together on the Web to create entirely new business models, and Web services will be the currency," Moss adds.

"Web services will legitimize the ASP [application service provider] market," says Peter Urban, a senior analyst at AMR Research Inc. in Boston. Urban continued that when popular applications become available via a Web-services model, ASPs will gain more traction and people will begin to see tangible results and benefits.

Standards and protocols

Moving forward, Web services will encompass just about any application available or imaginable delivered via the Web and standard protocols, especially UDDI (Universal Description, Discovery, and Integration), XML, SOAP (Simple Object Access Protocol), and WSDL (Web services description language).

A company offering a particular service describes it with WSDL then registers it with a directory, such as UDDI. A user in need of that service can then look it up in the directory and access it using XML and SOAP. For instance, a company conducting international commerce can access a service that approves global tariffs without having to own the tariff application or even subscribe to it in an ASP fashion.

There is no question that if Web services are to take off as smoothly as vendors hope, a significant chunk of the more than 20 million programmers in the world will have to write to UDDI, WSDL, XML, and SOAP. "The promise of Web services depends heavily on vendors supporting the standards," IBM's Hebner says. Although analysts say that more standards likely will develop, the four initial efforts will form the backbone of Web services. "These de facto standards appear to have momentum that [proves] they are going to be important," says Al Gillen, a research manager at market research firm International Data Corp., in Framingham, Mass.

The most important benefit of these standards is that client applications based on products from one vendor will be able to communicate with Web services even if they are based on software from another vendor, according to Barry Goffe, Microsoft's .NET product manager. "Web services will interoperate seamlessly, but probably not until about the third version," Gillen adds. "The first version will probably have a few glitches."

Web services are coming

The underlying principle of Web services is not entirely new. Dumb terminals and screenscrapers could be considered precursors of Web services. More recently, ASPs emerged to deliver applications across the Internet. "Hosted applications are basically a parallel strategy to Web services," says John Schachat, CTO of Networld Exchange, a San Diego-based online food services exchange. Schachat says, however, that mere application hosting does not encompass the potential for Web services.

Although much of the technology underlying modern Web services is still in development, Web services-like offerings are popping up in the market. OpenTable.com, in San Francisco, offers Web-based real-time restaurant reservations internationally and, in conjunction with partners, restaurant reviews and credit card processing. The service is even accessible (invisibly) via affiliated sites such as America Online.

"We're strong proponents of the Web-services philosophy," says Susan Lally, vice president of engineering at OpenTable. Perhaps that is because OpenTable is a natural: The company's business model essentially was a Web service before Web services were even born.

To move forward, OpenTable is supporting UDDI, XML, SOAP, and WSDL, with the long-term goal of making its service accessible from any device. Although client-side support is now available for WAP (Wireless Application Protocol)-enabled devices and PCs, Windows 2000-based OpenTable is using the standards to communicate with partners, including a very large credit card company that runs on Solaris, Oracle, and Java. "It's nice being able to write one interface and not have to even know what our partner's infrastructure is," Lally adds.

The road to El Dorado

Web services will change not only the nature of how applications are used, but also the way they are developed. "This is a radical shift for companies to expose their infrastructures to the Internet," says Shawn Willett, a principal analyst at Current Analysis Inc., based in Sterling, Va. Although a few of the Web-services vendors are trying to persuade the public that the development mind-set requires only slight alterations, developers and analysts alike are skeptical, if not downright disbelieving.

"People talk about XML and UDDI and the coming of seamless networks, but 90 percent of all applications were never componentized," says Robert Stimson, a vice president at Merrill Lynch and author of a recent report called Internet Distributed Services. "There will be a lot of adapters or gateways to make this happen. The componentization of all applications will be an evolution and take a lot longer than people think. There are no seamless XML interfaces right now. A lot of infrastructure work needs to be done."

The biggest issue with Web services may be translating current developer skills to forthcoming products, according to Martin Marshall, a managing director at Zona Research in Redwood City, Calif. "The burning question is whether Visual Basic is compatible with Visual Basic.NET, not whether Web services from different vendors are compatible with each other," Marshall says.

Microsoft in particular has had a difficult time making Visual Basic skill sets applicable to the Internet. "It took me a long time to move my [Visual Basic] skills and my FoxPro skills to the Internet. I think there is a psychological hump there," says Brian Kinkle, a senior project manager at Vega Application Development in Media, Pa.

Somewhere on the horizon

As is the case with most new technologies, Web services won't gain much traction overnight. Vendors such as Sun and Microsoft detailed road maps in which Web services-enabling products and technologies won't blossom for two to three years.

A report published in late January by Stamford, Conn.-based consultancy Gartner predicts that all leading e-business platforms will support at least the basic Web-services infrastructure by 2002 or by 2003. By 2003, more than 75 percent of Web services in production will be supported by Web-services infrastructures provided by IBM, Microsoft, or two or three other vendors, the report states.

When it comes to Web services, customers need to understand the technology, says Will Zachmann, a vice president at Meta Group in Stamford, Conn. "There is no way you'll have a clue about how Web services can be beneficial if you don't understand the technology," he says. "You better be able to understand this technology in order to run your business."

Any company considering the move should ask themselves what their business needs are, what infrastructure needs those business objectives create, and how they plan to communicate with Internet exchanges, according to Muntuck Yap, a principal in the Internet services division at Arthur Andersen, in Chicago.

"If you don't think about what a Web service can be, and design for that, you'll only be able to take advantage of about 25 [percent] to 50 percent of the benefit of these technologies," says David Story, vice president of engineering at Allegis, a San Francisco-based PRM (partner relationship management) vendor that provides its software in the form of a Web service to Fortune 500 companies.

Send comments to Senior Writer Tom Sullivan (tom_sullivan@infoworld.com). Martin LaMonica contributed to this report.

The .net effects

Web services will eventually be seamlessly interoperable with one another, but it is certain that developers and customers will still be faced with a perplexing platform choice.

At this point, analysts say, the choice is tough and may depend primarily on what architecture customers already have in place. "Microsoft has done the best job in winning the vaporware war at this point in putting forth a grand vision," says Al Gillen, a research manager at IDC in Framingham, Mass. "But it's not atypical for Microsoft to put a placeholder out in the market while they get their act together."

IBM and Sun Microsystems Inc., in particular, already have a number of the pieces in place for building Web services, although more are on the way -- whereas Microsoft's technologies are still in the beta-testing phase, particularly Visual Studio.NET and Whistler, the pending next-generation Windows OS; both are due later this year.

Gillen pointed out that Microsoft has some time to catch up technologywise while the de facto Web-services standards are maturing. Analysts agreed that Microsoft appears to have thought out its Web-services strategy more than the other vendors. And at the very least, Microsoft has done the most evangelizing on the subject. Some of Microsoft's competitors suggest that the Redmond, Wash., company's typical marketing blitzkrieg will benefit them as well as Microsoft.

"We won't spend as much as Microsoft on the brand itself, but we'll spend an inordinate amount of money on development," says Seth Pinkham, group manager for ASP marketing at Sun Microsystems, in Palo Alto, Calif. "The promotion that Microsoft is putting into .NET will go a long way toward making the world understand how business is moving to the Web."

Join the newsletter!

Error: Please check your email address.

More about AllegianceAllegisAmerica OnlineAMR ResearchAndersenAndersenArthur AndersenBowstreetGartnerHewlett-Packard AustraliaIBM AustraliaIDC AustraliaMartin LaMonicaMeta GroupMicrosoftOpenTableOracleSIRSun MicrosystemsTivoliUDDIZona Research

Show Comments