SLA Negotiation and Monitoring from Simulation Data

Abstract

The success of web services is changing the way in which software is designed, developed, and distributed. Services are in fact continuously re-designed and incrementally developed, released in heterogeneous and distributed environments. They are selected and integrated at runtime within external business processes. To ensure that a deployed service fulfils the QoS requirements, a Service Level Agreement (SLA) is associated to the service and allows to verify its behavior and to report cases where the service does not match the quality and performance requirements. After SLA negotiation, service monitoring is important to follow the evolution of its performance and to ensure that the service complies with the agreed SLA. In this dynamic context, this paper presents a methodology that aims to evaluate the performance of the service at early stages of the development process using simulation. The simulation data can be used first to negotiate the preliminary performance’s SLA of the service between the service providers and the customers looking for a given service, and later to monitor it. After the development, the SLA can be refined or updated to further match the real performance measured for the service by testing. This paper extends our previous works by using data obtained from our zero-knowledge approach, where neither the service code nor test-based the information on service execution times are available, to generate the SLA template and use it to negotiate the preliminary SLA and monitor it.