Web Services Journal recently asked its editorial advisory board for their answers on several questions about where Web services is going next year. The respondents include Paul Lipton, director and technology strategist, Computer Associates International, Inc.; Andy Astor, vice president for enterprise Web services, webMethods; Anne Thomas Manes, research director, Burton Group; Simeon Simeonov, principal, Polaris Venture Partners; Dave Chappell, vice president and chief technology evangelist, Sonic Software; and Bernhard Borges, senior technologist, IBM Business Consulting Services, EAI practice.

What Is Still Missing and Likely to Stay That Way in 2004?
Lipton: The larger challenges beyond Web service orchestration, such as business-level taxonomies and higher-level choreographies, will not be easily resolved in just one year. There still will not be widespread use of protocols and standards that function at higher business levels. We are still building the complex, distributed infrastructure that will be needed first. The higher-level semantics will start to come to fruition in 2005 and 2006.

Astor: Almost everything! Maybe we'll solve security next year, but I don't expect anything else to be truly resolved by 12/31/04.

Manes: For next year, I'm just looking forward to widespread deployment of SOAP and WSDL - let's walk before we run.

Simeonov: Customers still won't see the business case for significant SOA (service-oriented architecture) adoption. Problems will continue with SOA deployment and operations due to lack of standards and poor tools and management. Interoperable enterprise-grade Web services won't get much traction.

Chappell: What will still be missing is for an organization to be able to rely purely on Web services standards and interfaces to solve the problem of integrating on a global basis across departments and business units, and between partners. There's no big switch on the wall that someone can pull that will suddenly enable all applications within an organization to speak the same style of interface. IT is too busy just trying to "get stuff done" to worry about wholesale rip-and-replace of communications infrastructure. We will have to continue to live with the fact that disparate systems abound, each with its own means of connectivity and interfacing technologies. Technology leaders will need to innovate by bringing endpoint abstractions to the apps in ways that the apps can deal with it on their own terms, yet still provide a consistent manageable view to the "integration archtitect." Borges: We will still lack semantic standards and security.

What Issues Will Be Resolved in 2004, and How?
Lipton: Web services management will make major strides in 2004 with the completion of the OASIS WSDM (Web Services Distributed Management) specification and a broader understanding in IT shops that Web services truly need management, especially in a B2B setting. As Web services move from inside the corporate firewall to higher-value, mission-critical Web services, expect to see Web services management solutions that are fully SOAP aware from the major management vendors like CA.

Astor: See my earlier comments.

Manes: We've come a long way in 2003, so I expect that we'll achieve similar headway in 2004. We'll see broader adherence to the WS-I Basic Profile and use of Document/Literal services. We'll see widespread adoption of standards for advanced operational semantics, enabling security and reliability. I expect that we'll finalize the specs to support coordinated transactions, although adoption won't happen until 2005. I also think we'll make good headway with regards to policy definition, representation, and enforcement. Big issues that won't get resolved in 2004 include BPEL complexity, federated identity and trust relationships, and semantic integration.

Simeonov: We should be able to get SOAP 1.2 tooling and finally get a solid level of interoperability

Chappell: The term "Web Services" will no longer be heralded as the next big trend to solve everything. IT will be focused on integration more than ever, and the thought leaders in the industry will focus on articulating the architecture for enabling a highly distributed integration network based on standards. A new technology category, which is already being coined as the "enterprise service bus" and adopted by many vendors and analysts, will become a battleground for mindshare.

Borges: I see improvements in environments - partial and incremental resolution of architecture, design, (improved) best practices;
incremental resolution of BPEL implementations and management; and BPEIL engines - as the first tangible step toward business process as opposed to "mere" plumbing

What New Ideas Are Likely?
Lipton: There's bound to be an explosion, but I'm not going to hazard a guess here. That would spoil the fun!

Astor: The need for environments that support Web services from any environment, and that put the intelligence for this function into the network, rather than in the endpoints.

Manes: New ideas? I suspect we haven't thought of them yet.

Simeonov: There will be a focus on monitoring at the business level (real-time business intelligence leveraging SOA as opposed to the traditional BAM approaches). Maybe we shouldn't replace everything!

Chappell: It's very likely that new ideas around XML data storage, XML caching and aggregation, and intelligent routing of XML will emerge.

Borges: There will be a (continued) emphasis on SOA and process and on the service bus coupled with integrated semantic brokers

Will Business Process Take Off in 2004?
Lipton: Take off is a relative term. BPM (business process management) will certainly continue to grow. Specifications like WSBPEL and W3C Choreography will help that happen. As we move towards more complex composite Web services composed of myriad subordinate services, enterprises will find themselves compelled to use solutions that allow them to monitor and maintain service levels, quality of service, and more. The days of simple, unmanaged Web services are numbered for many applications.

Astor: Take off, yes. And by 12/31/04, I believe we'll know what the answer is going to be. But it won't be in place and finalized.

Manes: Many people already do business process - although today they typically use proprietary technologies. I expect that most vendors will convert their products over to BPEL in 2004. I just don't think we'll solve the complexity problem next year and I don't think BPEL will become a mainstream technology until we solve the complexity issue.

Simeonov: No, something like this doesn't "take off." The key impediment to this market is not the level of tooling but the state of business.

Chappell: End-to-end business process across departmental boundaries and business units will continue to be a challenge, and will continue to hinder its adoption. However, this will be top of mind for our readers, who will continually be looking for solutions.

Borges: Yes, but it will be limited by a lack of semantic standards and the inherent difficulty of (larger scale) ontology design