Enterprise computing is undergoing a radical shift: the move from hard-coded applications to lightweight services. To be sure, there is still code backing the services - the code implementing the applications -- but current IT dynamics emphasize interoperability as the least addressed and most lacking capability. The shift of focus to services underscores the growing importance of interoperability, which requires new IT architectures and federated governance. Service-oriented architecture (SOA) is the model that emerged to support these needs.

However SOA does not solve anything by itself. It requires adoption tools that support it, which are different from any that are designed to support client/server and mainframe infrastructures. Outward-facing services are distinctly different from traditional monolithic applications in the way they are packaged and used, which imposes delivery requirements that are different from those applied to application software. New relationships, represented by connections to services, are formed and dissolved every day, implying strict business-driven constraints for change management turnaround for services. Application logic is not nearly as volatile as are interconnections between loosely coupled services. Additional complications arise out of services linking two separate entities, which have independent IT governance structures with no top-down subordination. Most of the benefits that drive companies to adopt SOA simply cannot be realized when traditional ways of providing, monitoring and managing IT are applied to capabilities chosen to be packaged as services.

IDC believes that in the long term, Web services will remove or at least soften the constraints IT now places on business progress. IDC contends that making IT more responsive to business models will require a new applications architecture, one that encourages reuse, accommodates new technologies, enhances the value of the existing IT assets, is instantly adaptable, and is cost effective. Unfortunately, as Forrester responds: "SOA is hitting a wall in the companies that are most aggressive in its pursuit. Why? The profusion of service interfaces and their design artifacts, such as XML Schemas and Web Services Definition Language (WSDL), is out of control. Although some development or integration tools support limited storage and searching of this metadata, none are complete enough. The result: reduced reuse and a maintenance nightmare."

The problems referred to by Forrester are common for all but the most trivial interoperability projects. Their cause is the exceeding variability of data and interface design components, which cannot be easily harmonized due to varying engineering styles, programming and software design skills combined with heterogeneity of tool suites. UnitSpace helps to overcome this problem - as well as helping companies to quickly create and adapt Web services - by replacing the traditional procedural approach to creating Web services from scratch like other software components with a new declarative approach, which radically improves manageability of SOA and standardizes service interfaces across multiple providers.

The procedural approach to the creation of Web services suffers from several inherent drawbacks:

• information systems' internal data models are not interoperability-ready and thus limit service utility and constrain reuse;
• lack of separation between interoperability architect, Web services programmer and application developer results in poorly designed service interfaces and unsatisfactory implementations;
• inadequate attention given to security, performance and manageability;
• similar functional requirements lead to multiple implementations of functionally equivalent Web services in different systems.

Overview: The Solution