Automatising Mashup of Cloud Services with QoS Requirements

Service mashups represent an appealing business opportunity for com- panies since value added applications can be provided to fulfill clients' needs by integrating their services with the ones available on the Internet accessible accord- ing to standard Web Services technologies. Clients' needs are usually expressed in terms of a required functionality that can be obtained as a mashup application, to- gether with specified QoS requirements referring to non-functional characteristics of the application, such as price, time, reliability. In order to make this opportu- nity a reality, mechanisms allowing for automatic selection and composition of ser- vices are necessary to avoid human intervention in the composition process. Here, a framework for automatic mashup of Cloud services taking into account QoS users' preferences, is presented. It relies on both AI planning techniques for automatic ser- vice composition, and software agent negotiation to select a composition that meets the specified QoS preferences. It allows for a dynamic QoS-based mashup of ser- vices since the QoS values provided for the single services in the composition are not fixed, but they could vary according to the providers' strategy. The proposed approach can be applied when services are provided in the context of a competitive market of service providers.