The e-business revolution prompted companies to abandon their company-centric focus and broaden their definition of interactions with customers and business partners; as a result, many companies have developed interconnecting business processes – and applications – over the Web. The latest technology, Web services, defines a common set of standards that facilitates bridging incompatible systems and databases, thereby invading the traditional domain of EAI (enterprise application integration).
How will Web services affect EAI? Respondents to the 2002 InfoWorld Application Integration Survey have attitudes toward the future of traditional EAI solutions that vary from outright dismissal to cautious confusion. It’s possible to define Web services as a set of supertools for application integration that will replace legacy EAI solutions. More than 50 per cent of survey respondents believe that Web services will eliminate or significantly reduce the need for EAI tools; but 36 per cent predict null or marginal changes.
Web services can solve almost any integration problem at its root, so its adoption could make traditional EAI tools redundant. But significant differences between the two approaches suggest caution when declaring traditional EAI dead.
One significant difference between the two technologies is that converting an application to Web services is a permanent cure for platform compatibility issues; traditional EAI tools merely create a bridge between applications, which solves the problem at hand but leaves the way open for future integration woes. Consequently, adopting Web services has a steep initial cost but produces smoother and less expensive future integrations.
Obviously, the difference in integration cost and efforts becomes more significant as the number of linked applications increase. CTOs should select Web services for integration projects that target multiple applications, while maintaining the traditional approach for projects with a smaller-scale impact. For example, a customer database, a critical resource for most companies that is often shared by most departments and even by outside partners such as OEMs and VARs, is a likely candidate for Web services, whereas the traditional approach would be more cost-effective for a payroll application, which normally has fewer demands for integrations.
When asked why they would adopt Web services, 25 per cent of respondents (the greatest single answer) pointed to CRM, a likely candidate for multiple integration projects. And 51 per cent expect that Web services will make EAI projects less expensive.
If results of the Application Integration Survey are any indication, companies will adopt Web services en masse for an unprecedented and long-awaited freedom from compatibility problems. In fact, 67 per cent of participants are investigating Web services because they can offer a complete or partial solution to their integration problems. But for 22 per cent, integration issues have nothing to do with their interest in Web services; for them, the main appeal of Web services is its capability of facilitating e-business.
Even if they don’t see Web services as an integration cure-all, 56 per cent of survey participants expect it to make integration projects more palatable and are likely to result in increasing the number of EAI projects they pursue. Users expect vendors to deploy Web services to facilitate integrating monolithic suites with third-party software.
IT managers welcome Web services as a unified application platform that will finally remove the burden of costly and time-consuming technical re-engineering from EAI projects. Most survey respondents expect that adopting Web services will reduce or eliminate dependency on conventional EAI tools. But this high hope is likely to be dashed, at least in the immediate future.
Vendors of traditional EAI tools haven’t lost their hold in the enterprise; many respondents seem to not be sure about what to do this year. Fifty-seven per cent of survey respondents indicated that they currently use one or more EAI brokers from major vendors, including BEA Systems, IBM, Microsoft, Oracle, Sun, and webMethods. But 52 per cent stated that they are not using any EAI broker. And 47 per cent of our respondents declared plans to buy EAI brokers this year.
A similar trend is behind the purchasing of EAI product suites: 52 per cent of survey participants have installed at least one or more products; 42 per cent of them are focusing on major vendors, such as Attachmate, Data Junction, Hummingbird, IBM-Neon, NetManage, and WRQ. But 45 per cent of participants are not using any EAI suite. Thirty-six per cent of respondents estimated that during the next 12 months they will purchase one or more EAI suites.
Clearly, many IT leaders see Web services as a long-term opportunity to standardize their applications. In the near term, they will continue to use – and even buy more – traditional EAI tools. Even companies that adopt Web services will still deploy EAI solutions to solve specific business problems such as integrating legacy back-office suites. The future traditional EAI appears to be limited to islands of information, restricted to company or departmental boundaries. Nevertheless, IT leaders will need traditional EAI tools for as long as software vendors offer legacy, non-Web services suites.
IT leaders are going to standardize their applications on Web services, and they expect vendors of EAI solutions to do the same. In the future, EAI vendors will offer integration services rather than packaged software, making their solutions an integral part of the Web services family. Paradoxically, to survive, EAI solutions will have to assimilate their nemesis.
THE BOTTOM LINE
The Future of EAI
Executive Summary: The standards-based architecture of Web services can offer less expensive and simplified integration of business processes, whereas traditional EAI tools can address specific tasks as an alternative or complement to Web services.
Test Center Perspective: CTOs should identify Web services candidates among EAI projects according to business relevance and scope. Priority should go to integration products that adopt technologies consistent with Web services standards.
