Myths around Web Services

Web services and the technology surrounding them have become the dominant trend in the electronic commerce arena. XML, SOAP, UDDI, and WSDL, as the foundation of Web services, are all attracting considerable attention as potential bridges between heterogeneous systems distributed acrass the Internet. The assumption seems to be that soon most applications will speak and understand XML, that all systems will support SOAP, that everybody will advertise their services in UDDI registers, and that all services will be described in WSDL. Once that stage is reached, application integration and business to business (B2B) e-commerce will be straightforward. Unfortunately, Web services are only the next step in the natural evolution of middleware. Therefore, by design, Web services are evolutionary rather than revolutionary. The most basic form of middleware are RPC engines. When such engines become transactional, they become TP-Monitors. Once object orientation aspects are inc1uded, TP-Monitors evolve into Object Monitors. Message oriented middleware (MOM) also originated fram TP-Monitors since persistent queuing was a feature of many TP-Monitors until they became systems on their own. In fact, many· MOM platforms are RPC based. Web services are, primarily, an extension to middleware platforms to allow them to interact across the Internet. Only from this perspective do many of the developments in the Web services arena make sense, e.g., that one of the first protocols to be wrapped as SOAP messages was RPC (and, at the time of writing, almost the only one to have been completely specified). Of course, it is possible that Web services will trigger a radical change in the way we think about middleware, application integration or the way we use the Internet. In that case, Web services will evolve into something yet unforeseen. At this stage, however, this has yet to happen. In reality, and precisely because they were created with that purpose in mind, Web services are used today almost exc1usively for conventional enterprise application integration (which may or may not happen in a B2B setting). It is this experience as an extension to middleware platforms that will define and shape Web services in the short and medium termo There are nevertheless many proposals that take Web services well beyond their current capabilities: semantic web, dynamic marketplaces, automatic generation of B2B applications, seamless integration of IT infrastructures from different corporations, etc. These proposals are the basis for presenting Web services as revolutionary, rather than evolutionary. Such speculations are the province of long term research but they tend to ignore the exact nature of Web services and the underlying technology. Many of these ideas also ignore the current limitations of existing middleware platforms although most of these limitations appear again at the Web service level. In this regard, Web services have to a certain extent become an outlet for ideas that proved impractical in the