TechWeb

Look Past Products To Web Services' True Promise

Aug 25, 2002 (08:08 PM EDT)

Read the Original Article at http://www.informationweek.com/news/showArticle.jhtml?articleID=6502983


With nearly every enterprise software vendor pledging support for Web services recently, much of the buzz surrounding them has centered on the technology. Product features and functionality certainly count, but they don't embody the real importance of Web services. Instead, the technology is poised to make its greatest contributions as an enabler in the evolution from conventional, distributed computing to service-centric computing.

Service-centric computing provides a high-level model for how applications can be assembled from a distributed network of constituent parts, and Web services contribute the nuts-and-bolts details for technical implementation. The service-centric model also creates a new ecosystem of interrelated suppliers and consumers of not just technology but information. It does so by providing access to distributed services, which themselves serve as information providers. The ecosystem being created around this web of information providers and consumers includes software vendors, application service providers, publishers, exchanges, and standards bodies.

Suppliers of technology components such as server software, development toolkits, and service frameworks have gotten the most attention as the business-technology world has turned to Web services. Now it's time to turn to service-centric computing's ability to catalyze fundamental changes in the way software services can map to real-world functions and processes.

In the real world, business relationships and processes cut across groups inside companies, as well as across companies, with individuals and automated systems making contributions along the way. Efficiencies are introduced into the system when the friction between these interactions is reduced. Service-centric computing can reproduce these interdependencies as well as optimize the interactions.

It's important to separate the tool from the implementation in service-centric computing. To put that concept in perspective, think back to the avalanche of hype that accompanied the public's discovery of the Internet. Money flowed into any technology that billed itself as Internet-related, with the implicit understanding that the simple act of having an online presence would be enough to breathe new life into old-economy businesses.

One fatal flaw seems obvious in retrospect: Focusing on enabling technology as the end game, rather than as a means to an end, proved to be a critical mistake. Given the carnage in the technology market, this distinction seems all the more relevant. In the late 1990s, many businesses made investments in technology without fully justifying the cost versus tangible return on investment. One result: an explosion of shelfware that never helped users achieve their goals.

The fact is that the Internet never made anyone rich, never increased a company's bottom line, never made anyone's life easier. What the Internet does do is provide a valuable tool that lets businesses and individuals benefit from squeezing inefficiencies out of their processes. Just as a Steinway piano won't turn an ordinary person into a virtuoso, new technology won't let an IT department with a lackluster plan achieve significant benefits.

In a similar way, even though some Web-services implementations are available from enterprise-software vendors and others are on the drawing board, buyers won't benefit without road maps that clearly define how this technology will apply to and help automate their existing business processes.

The first question to answer is whether Web services can address an existing pain point. For most businesses, this answer is clearly yes. But for smaller shops, the benefits may not be as obvious. And for some larger companies, full-blown implementations of enterprise application integration technologies may be more practical. Where security is critical, it makes sense to hold off on Web services until more-comprehensive standards are put in place.

Yet once you've identified specific business problems where Web services make sense, plan your implementation road map in a series of discrete phases and use earlier phases as proof points for more-extensive architectures. Learning the limitations of any new technology through a series of small, behind-the-firewall implementations is a prudent step to take before deciding whether to commit serious resources.




Aside from their ability to serve as a catalyst for service-centric computing, Web services are poised to transform the way we buy and use shrink-wrapped applications. For example, computer users may no longer need to buy desktop software in retail locations. They'll merely download snippets of code that reference application modules. These modules may be installed locally or could just as easily reside on remote systems as services. The economic impact of this approach will be the gradual elimination of the retail-reseller link in the supply chain, which will result in greater profits for the creators of the applications and lower prices for consumers.

In service-centric computing, software-licensing models and software vendors also will be reshaped. Most software is licensed for a set number of machines, for an indefinite period--a model that has considerable implications for a software company's revenue stream. In this model, the release of a new iteration of a product creates demand, and the vendor's resulting revenue typically spikes, as do the associated marketing expenses. Over time, sales slow and revenue from this product drops. A service-centric model evens out these peaks and valleys, which leads to a more predictable revenue stream and greater visibility in sales forecasting.

From a user's perspective, the service-centric model allows software to be used as needed for a fixed period or on a pay-per-use basis. One positive outcome of this arrangement for IT managers is that users can receive updates that install themselves without the need for a technician to be sent to every desktop. Also, software delivered as a service doesn't require the elaborate anti-piracy systems found in current products, since services can be easily shut off if fraudulent use is suspected. This frees users to potentially utilize a software service on any machine at any time, creating true device independence.

The payoff for software manufacturers is that they can profit from more-predictable revenue streams, which are no longer so closely tied to upgrade cycles. The big advantages for users: reduced administration costs and automatic software upgrades.

The term Web services, as these examples illustrate, is a bit of a misnomer. There's nothing about Web services or service-centric computing that requires the use of a Web browser. In fact, each day the notion of the browser becomes more dated. The emergence of new Internet-attached devices such as set-top boxes, kiosks, and game consoles, combined with the decline in PC sales, underscores the fact that browser wars will become somewhat irrelevant in future generations of client technology. The impact for users is that they will be able to pick and choose how they access services from a variety of devices rather than being tethered to PCs.

While the previous examples represent point-to-point communications models, service-centric computing can be extended to encompass interrelated networks of services and resources. These peer-to-peer networks, where client computers communicate directly with one another, really comprise nothing more than groups of computers, each exposing one or more services. The truly democratizing effect of service-centric computing in this context is the fact that no central server is needed to control interactions among users. All individuals on the network can provide and consume resources as they see fit. Of course, the logical next step for this kind of technology is the creation of grid networks, where resources can be provided dynamically, reducing computing power and service provisioning to the commoditized level of power-generation and distribution systems.

However, the success and adoption rate of new technology ultimately depends on the value it provides to end users, and the bottom line for Web services is no different. Because Web services will enable fundamental changes in the way software is packaged, delivered, and consumed, it's key to have a solid road map to make sure the impact is positive.

With Web services, as with any software project, your chances of success hinge on basic but essential steps such as performing a thorough needs assessment, gaining a good understanding of the technology and the vendor landscape, and using a pilot program before undertaking large implementations. Once those fundamentals are covered, chances are that Web services will result in greater efficiencies for your business. And that's what will transform the technology from today's hype to tomorrow's reality.

David Homan is a principal analyst with Doculabs, a research and consulting firm. You can reach him at info@doculabs.com.