There is an interesting article in this months IEEE Computer Magazine on Commitment-Based Service Oriented Architecture. The authors, Munindar P. Singh, Amit K. Chopra, and Nirmit Desai, put forth that existing implementations of service oriented architecture are low-level abstractions instead of business services. They propose a new Commitment-Based SOA where the components are true business services and the connectors are patterns.
As proof of this claim the authors provide the hypothetical example of a purchasing of an item that combines ordering, paying, and shipping business services. The current approaches such as Business Process Modeling Notation, Business Process Execution Language, and Choreography Description Language all emphasize controll and data flow. In contrast to these the CSOA “gives primacy to the business meanings of service engagements, which are captured through the participants’ commitments to one another.” Commitments include three agents: debtors, creditors, and context. The autors claim benefits of the CSOA are 1) Services can be changed or modified and judged to be correct as long as the commitments are not violated. Commitments thus support business-level compliance and don’tdictate specific operationalizations. 2) Commitment-based specifications explicitly reflect business requirements. Without this business meaning, there is no basis to establish the validity of reuse or composition.