IntroductionService discovery is the matching of the needs of a service requestor with the offerings of service providers. It is currently functionally focused (i.e. concentrating on what the service does), and occurs in what can only be considered a volatile and heterogeneous environment. Bring to that equation the area of web services, with its various specifications, and service requestors are no better equipped to efficiently discover both traditional and web services. Subsequently, the interactions that occur with service providers are drawn-out, challenging affairs that leave most service requestors wondering where the word "service" applies. Issues that arise for service requestors include the inability to compare services objectively, insufficient details about either or both the functional and the non-functional properties of services (e.g. price, payment, temporal and locative availability), no common language for the description of services (i.e. different service providers use or misuse service related terms), and a lack of mechanisms to assist with effective filtering of relevant services. So how do you find a service that meets your needs? MotivationWe believe that services need to be well-described. The descriptions must be semantically rich and they must have structure. We want to exploit that structure to enhance the discovery of services. To be clear, we view the description of a service as consisting of two halves. The first provides a description of the functionality of the service (i.e. what it does), the latter provides a description of the non-functional properties of the service (i.e. anything that exhibits a constraint over the functionality). An example of a non-functional property of a service is its temporal availability. One dimension of temporal availability for services can be considered the constraining of when a service can be accessed by a service requestor. All these factors, but in particular the lack of description of non-functional service properties, hinder the support for sophisticated service discovery, service selection, automated service negotiation and dynamic service substitution. Our intent is to reduce the "fulfillment gap" that is generated by incomplete information and an internal understanding gap caused by a communication disconnect [37]. The primary motivation behind our research is to utilise this semantic richness of service descriptions to enhance service discovery and substitution. We are interested in the description of these constraints that are exhibited over services. We consider that these constraints may be imposed by the service provider or by the environment within which the service operates. Our aim is to develop an abstract syntax for a language. As such it exhibits no commitment to items such as keywords or ordering. We do not envisage that our abstract syntax will be capable of describing all services. We prefer not to distinguish between electronic and traditional services. This is because we may request them electronically and receive them traditionally, or vice-versa. We feel that the description of services must cater for both traditional and electronic services.
|