Browse by year:
March - 2003 - issue > Cover Feature
Goodbye Application Integration, Hello Standards-Based Integration
Uttam Narsu
Thursday, February 27, 2003
As commoditization through Web Services looms large, application integration vendors are completing their transformation to BPI/BPM vendors. Will it be too late?

APPLICATION INTEGRATION (EAI) IS OFTEN touted to be the “killer app” for Web Services, which it isn’t. Web Services don’t need one, as judged by Giga surveys that show sharp increases in Web Services’ intentions to spend, across a variety of use cases, despite the economic downturn. Yet Web Services are highly immature today, while the classical EAI solutions are overflowing with rich functionality. Therein lie the seeds of the problem. Fewer than 70% of application integration users use less than 30% of the features of an EAI solution. Moreover, users of current EAI vendor solutions represent a small fraction of application integration spending by enterprises: the majority of point to point integration is done using file transfers, remote procedure calls, and other hand-coding techniques—not EAI products. EAI products remain too expensive for those users.

The need to serve their current high-end customers has forced most of the application integration solutions to complete a transformation begun many years ago, before the threat of Web Services emerged. At the end of the 1990s, application integration vendors incorporated business process integration capability (BPI), targeting real-time data and process exchanges over the Internet, and in the last eighteen months, they have added business process management (BPM) functionality to support end-to-end value chain integration. This investment at the high-end of the market has precipitated a classic innovator’s dilemma, where the vendors’ rich investments in sustaining technologies, to use Clayton Christensen’s terms, have focused on the needs of current customers at the expense of new customers. All that was necessary to complete the picture was a disruptive innovation, and that emerged through the guise of standards-based integration (Web Services, JMS, JCA, and JMX). As the threat became clear, platform vendors signaled their intention to be the locus of enterprise integration spending by acquiring EAI vendors.

More tellingly, by simultaneously investing in application integration capability and disruptive standards-based integration technology, the platform vendors have won the marketing battle. Today, 60% of respondents in a recent Giga survey expect to get their Web Services support from their primary application platform vendor, as opposed to just 17% from their application integration vendor. 47% of these same respondents pursue Web Services for application construction, 25% for application integration, and the remaining for the sales value chain. This interest arises despite Web Services’ immaturity: of the eventual 10 to 15 Web Services standards, only three (SOAP, WSDL and UDDI) have widespread agreement, and the lack of security, reliability and transaction support are the three biggest inhibitors to widespread enterprise adoption. This interest also spans a range of enterprises, from financial services and banking, to insurance, manufacturing and, intriguingly, the government at multiple levels. Web Services are clearly moving beyond the early adopters and innovators to early mainstream adoption.

Services Blur the Line Between Applications and Infrastructure
While there’s no doubt that Web Services are immature for all but the simplest integration use cases, there’s great faith in their potential because they support the higher level macro-trend in enterprises: a move to a service orientation. One of the arguments Giga clients make in favor of using Web Services both for application construction and application integration, is the harmonization and simplification in the infrastructure they use across custom and packaged applications, and the connections between them. As applications are deconstructed into less monolithic structures, moving toward being composed of smaller components and services, the lines between infrastructure and applications blur. Once the deconstruction is underway, aggregation can begin, constructing services composed as aggregations of other components or services. The point of all this work is to enable greater rates of change to the IT portfolio, in a shorter time, thus realizing the vision of IT for value creation, not as a cost center to be reduced.
How Will the Wolves Survive?
To survive in this world, the application integration vendors have focused on the business process level, the macro level, and the level most connected to the business side. While all these Web Services innovations will take place at the business process level, business processes differ greatly between a telecom provider and a bank. To succeed at this level will require addressing more business-specific elements. Increasingly, the risks around the survival of application integration vendors will be vertical-specific: can they establish a foothold in specific vertical markets that are large enough to sustain them?

Certainly, last quarter’s performance for most of the application integration vendors was surprisingly strong, which reflects their ability to do much more than the simple point-point integration typically found in Web Services solutions, or the pub/sub found in JMS. There is also the “trading partner drag effect” which decreases the willingness of a partner to change its infrastructure. Depending on the size of the ecosystem around an EAI vendor, this drag will result in strong maintenance revenue, if nothing else. Yet this drag effect works both ways: standards-based integration offers the lure of a larger potential community that is more independent of the infrastructure each participant has, rather than the smaller communities arrayed around each of the EAI vendors. Even as the EAI vendors embrace standards, they have to compete with cheaper alternatives available from a bigger platform player. And that will be the tolling of the bell.

While many vendors may survive in certain vertical niches, they will find that the proportion of each sale devoted to services will sharply rise, simultaneously making them more dependent and competing with their systems integration partners.

Enterprises must develop a multi-faceted strategy to handle their integration investments. This application and infrastructure strategy must increasingly focus on integration as a necessity for new application development, emphasizing architectural best practices to better transition IT away from application-orientation to service-orientation. As a first step, an approach based on application portfolio management (APM) as a means of gaining control of application inventory, minimizing maintenance and enhancement costs, promoting expedient application understanding, and aligning applications to their business owners is highly beneficial. Doing so will place the role of standards-based integration, including Web Services, firmly in the business context.

Uttam Narsu, a Vice President with Giga Information Group, covers a broad range of enterprise application development areas. His current research focus is on XML, Web Services and agile development methodologies and processes. Uttam has spoken at numerous industry conferences, including eleven in 2002 alone. Narsu has more than 18 years of experience as an analyst, architect, software developer, trainer and consultant.

Share on LinkedIn