Where will Web services be deployed? – Part 2

Viewpoint: July 20, 2002

In my last entry, I referred to two sources of confusion about the deployment of Web services technology. At that time, I tackled one of them – the notion that Web services are all about dynamic composition of applications. Now, I want to address the second source of confusion – the view that Web services will first be deployed within the enterprise and only later will bridge across the firewall.

On its face, this view is very plausible. After all, Web services technology is still at an early stage of development. It has many limitations. In particular, there is very little experience with approaches to maintaining security while using Web services technology. Beyond the firewall lies a dark and fearful area, filled with unknown threats and nameless hackers.

The reality is quite different (whoever said that reality was plausible?). The experiences with early adoption of Web services in large enterprises suggest that early adoption is proceeding along two parallel tracks. On one track, IT organizations are launching experiments with the implementation of Web services technology. These experiments are carefully contained so that they do not disrupt business operations – they are more demo projects or prototypes, rather than commercial production systems. They are primarily designed to build knowledge and skills within the IT organization regarding the capabilities and limitations of Web services technology. It is this track that technology analysts tend to focus on. After all, their primary contacts within the enterprise are with IT executives. And, with regard to this track, the technology analysts are right – the experiments are all carefully guarded behind the enterprise firewall.

All well and good. But if this is the primary vehicle for Web services adoption, it will be a long road ahead. Experiments take time to conduct and evaluate. Once evaluations are in, IT executives must identify promising areas for commercial deployment and gain the support of appropriate line executives. In the meantime, vendors will face the discouraging prospect of extended sales cycles and minimal purchases of technology.

But, wait a minute. There is a second track that is largely being pursued under the radar of technology analysts. This second track consists of business line executives facing real, near-term business issues – most commonly, these days, how to get more operating (expense and asset) savings quickly. In particular, these are line executives operating at the “edge” of the enterprise – responsible for activities that involve frequent interactions with business partners, typically either with suppliers and logistics providers in the supply chain or customers and channel partners in the sales, marketing and customer support functions. They are frustrated because a lot of the inefficiency that could yield near-term operating savings is concentrated in those interactions with business partners. Supported by sneaker net or swivel chair integration today, these interactions would consume far less resource if they could be automated by connecting applications and databases together and taking staff out of the middle.

But there’s a problem. Previous generations of technology are remarkably ill-suited for the diversity and complexity of automated connections that span across multiple enterprises. Electronic Data Interchange (EDI) connections, the old staple, are expensive, require significant technology expertise within each of the business partners and are very difficult to modify once established. Enterprise Application Integration (EAI) technologies are largely designed to solve integration issues within the enterprise, are expensive to implement and are relatively inflexible (although all of these drawbacks are being addressed over time). Bottom line: line executives face business problems that previous generations of technology simply can’t handle.

The response? Some brave (and desperate) line managers are proceeding to address these problems with Web services technology. Somehow, somewhere, they heard about the technology, did some research on it and found some people with the skills required to implement the technology. Often, they are doing this over the active resistance of their own IT organizations. But they are reaping tangible business benefits with remarkably quick paybacks and learning a lot about new possibilities in the process (for some early examples of business impact from Web services adoption, see my working paper “Break on Through to the Other Side: A Missing Link in Redefining the Enterprise (PDF)”).

In many respects, history is repeating itself. Or, as someone once said, “history doesn’t repeat itself, but it certainly does rhyme.” Those of you old enough to remember the early deployments of minicomputers will recall that line executives latched on to this technology as a way to gain independence from the priesthood of the mainframe. The early generation of personal computers snuck into the enterprise as “miscellaneous business expense” by line executives lusting for the spreadsheet application that could be run on this remarkable device.

Once again, line executives are plunging ahead, driven by business needs that can’t be addressed in any other way. Looking at it from a technology vendor perspective, this track is a good news/bad news story. First, the good news: large-scale deployments of the technology (often encompassing hundreds or even thousands of business partners) are being made today in real production environments and yielding quantifiable (and significant) business benefits. This should significantly accelerate the adoption of the technology as referenceable customer case studies multiply. In the near-term, significant purchases of technology are on the table.

Now, the bad news: there aren’t yet a lot of these deployments. Finding line executives with the courage to pioneer the adoption of a new technology isn’t easy – they are dispersed throughout large enterprises (but, remember, they tend to be concentrated in the edge functions like procurement, sales and customer support). Technology vendors used to selling to IT executives and relying upon existing relationships with these executives now face the daunting challenge of finding and effectively selling to a very different kind of decision-maker. Much greater depth in the business context will be essential. The ability to minimize techno-jargon will be equally critical. It won’t be easy, but for those who succeed they will enjoy both large near-term purchases of technology and the development of referenceable accounts that can turbo-charge future purchases of technology.

So, in many respects, the real action in terms of early adoption is occurring at the edge of the enterprise. It is here that connections are being deployed across applications and databases to remove operating inefficiencies with business partners. These connections venture out beyond the firewall, despite all the risks (and there are ways to manage these risks), because it is precisely here that existing technologies are least helpful. At least within the firewall, executives can look to EAI and other technology solutions, for all their limitations, as a viable option. When they reach the edge of the enterprise and peer across that firewall, they often have no viable options. The technology is in fact being deployed where it is uniquely able to solve real business problems.

That is not to say that Web services technology will remain at the edge of the enterprise. As the business benefits of the technology become broadly visible in these edge functions, the technology will get sucked into all manner of integration initiatives within the enterprise. Ultimately, it will redefine the technology architecture of the enterprise.

So the analysts have got it only partly right. Some limited implementation of Web services technology is occurring within the firewall, but mostly in the form of experiments. The real action in the near-term, both in terms of business impact and vendor opportunity, is occurring elsewhere. Look to the edge. Find the firewall and then look across it to find executives tearing their hair out over process inefficiencies massed on the boundaries of the enterprise. This is where the early successes, and long-term opportunities, for Web services technology are being shaped.