One of the most enjoyable aspects of my job at Progress is my daily contact with standards organizations like the TM Forum and OASIS. I work with many other industry-supported contributors to develop and extend the frameworks and guidelines that address the challenges they face in application development and systems integration.
Within the last month the TM Forum has issued for member evaluation new versions of three NGOSS frameworks: the Business Process Framework (eTOM) 7.5, the Applications Map (TAM) 3.0, and the Information Framework (SID) 8.0. These releases continue the evolution of the frameworks in both breadth and depth, and also support the upcoming release of the new NGOSS Contracts guidebook.
NGOSS Contracts extend the reach of the NGOSS frameworks into the system and implementation views. These contracts provide standards for application interaction that complement the existing SOA view of services and operations by defining the bigger picture within which those services run, such as conditions and constraints, message exchange patterns, and service level agreements. NGOSS Contracts also represent a meeting place between top-down service design based on the interplay of process, data and application models, and bottom-up service design based on real-world experience with successful service function and granularity.
Having seen firsthand both the promises and disappointments of client-server, CORBA and now SOA, I believe that contracts like those promoted by the TM Forum will make a critical difference in the long-term success of SOA infrastructure compared to earlier architectures, because contracts can cleanly separate the process implementation from the data model. This opens up the possibility of using a data integration layer, along with a robust business process management layer, to achieve real separation of business process from application interface details. The process layer can then select from alternative applications at a fine-grained level without the need for application-specific process changes, allowing an enterprise to reduce or eliminate expensive customization and to encourage competition between application vendors. Successful large-scale OSS transformation projects have demonstrated the technical and economic validity of this approach.
The importance of the data layer is also emphasized in a new white paper by Dave Hollander, a co-inventor of XML and a veteran of changing data standards. Dave examines the disparity between SOA's success in large-scale service interconnection, and its failure to address operational differences between applications, resulting in disappointment and failed projects. He concludes that the common model architecture is the key to achieving that semantic integration. Dave's paper, Common Models in SOA: Tackling the Data Integration Problem, will be available soon and I'll send a link to it when it's finalized.