Towards a Precise Understanding of Service Properties

We seek to develop an abstract notation capable of representing services. We view the need to describe a service as analogous to the need to label goods or products in a supermarket. A product label provides a brief summary of the good to which it is attached. Prospective buyers can use this information, together with the price, to make a rational purchasing decision. Product labelling occurs for the safety and benefit of purchasers. We believe the same accurate labelling (or description) should be provided for the benefit of service consumers. This research is being is driven by a series of problems currently facing the electronic services domain. These problems include:

• The ability to capture all the functional and non-functional properties of services.
• The ability to optimise composed services. In particular, to dynamically substitute one service with another satisfactory service (both functionally and non-functionally).
• The ability to validate a service. By allowing a trusted third party to invoke the service and compare the invocation of the service with the description of the service.
• The ability to discover services based on non-functional properties (e.g. temporal availability). Once developed, our abstract notation will be capable of addressing these problems facing electronic services.

The service description language will be capable of expressing properties such as pricing strategies, pricing factors, charging styles, settlement models/contracts, spatial and temporal availability, channels (request and delivery), service rights, service quality, security and trust. A direct result of this language will be the increased level of automated support for services, thereby introducing new levels of service dynamism and flexibility.

Status: Expected Completed Q4/2004