Web services are already changing the integration model for enterprise applications, and they offer even greater promise for the future. It's not going to happen overnight, but there is an ever-increasing focus on incorporating Web services into new applications and existing infrastructure in order to achieve standards-based information exchange. This expanding universe of Web services will allow not just J2EE applications, but all applications, to interact and exchange information freely, and permit development to proceed unencumbered by proprietary communications constraints.

Both J2EE and the Microsoft .NET framework have pushed the edge of the Web services envelope. Both platforms have already adopted Web services as their future integration model, demonstrating that Web services will become even more widely accepted and soon emerge as the dominant integration method.

Currently, J2EE is the prevailing enterprise development environment. More and more enterprises are already starting to expose their massive installed base of J2EE applications as Web services. The J2EE platform is now extremely powerful and broad in its capabilities, integrating features for both the server and client sides to enable existing enterprise Java components to be exposed as Web services and Java clients to invoke them. And more Java advancements are in the works.

The Java community is already at work on J2EE 1.5. The next release of this powerful platform incorporates new standards and language features from the recently announced JDK 5.0 and makes it even easier and more productive to create Web services. The JDK 5.0 language extensions support more productive Java development and deployment, enabling the Java code to be annotated with higher-level business process and other information. These improvements allow even higher-level abstractions to be built on top of Web services to solve real business problems, mapping directly onto solutions that come from business domain experts and enabling Java to more naturally support business process execution language (BPEL) environments.

As J2EE has advanced, so has the maturity and number of .NET applications. Soon Web services developed for these platforms will be able to communicate freely and without restrictions based on their platform. Development shops will be able to take advantage of each platform depending on their specific business needs, without having to worry as much about integration and interoperability issues.

Even with all this progress, Web services are still at the beginning of their evolution. There will be continual pressure to evolve specifications for both .NET and J2EE and create new specifications that can help IT teams better manage these complex transactions.

Security, reliability, and performance are the three primary areas where there is the most immediate need for advancing Web services. Some in the industry believe comprehensive Web services security, the kind to bet a business on, is nowhere near where they would like to see it. And reliability, especially of third-party Web services, is still elusive.

Following the natural progression of most computing paradigms, Web services security will likely be the first area of concentration for improvement, with high availability of Web services not far behind. However, until Web services and the platforms on which they are built advance in these areas, enterprises may decide to stick with the older integration model of components, especially enterprises that demand extreme levels of performance, security and reliability.

Much work is already underway by vendors, the open source community, and standards organizations to address these challenges. In the meantime, enterprises can manage system performance and employ tools to help them achieve better performance of Web services today.

Tools available from Borland and other companies can already help manage and measure the performance of Web services, and help tame some of the complexities of developing a service-oriented architecture (SOA). For example, some of today's tools can simplify everything from modeling high-level views of enterprise SOA implementations, to applying best practices and successful patterns to implementations, to auditing software for security holes before the software is deployed.

Eventually, security and reliability, as well as performance, will be addressed at a more functional level within the technology platform itself. Vendors such as Borland, Sun, BEA, and IBM are already collaborating to push the technology forward. Yet as powerful as these new platforms promise to be, this power will still come at a cost. It's the cost of complexity - today's systems are very robust but very complex.

The level of complexity involved in SOAs today requires enterprises to look beyond just platform advancements and more into the way they manage their software and IT systems. Businesses need to demand the same level of rigor and predictability in how they develop and deliver software as they do in other critical aspects of their business. Today it's too chaotic. There must be a higher level of predictability in building SOAs, more visibility into the risks and costs involved, the impact on business and development teams, and a better understanding of the benefits and return on investment a business hopes to achieve with these plans.

This is an issue no matter what platform an enterprise relies upon for Web services.