Self-Adaptive Fuzzy Architecture to Predict and Decrease e-Contract Violations

Abstract

Web service monitoring is important for consumers and providers. For the consumer, it is important to know if the provider is respecting the established electronic contract. Providers must satisfy functional and non-functional features as required by consumers. This paper proposes a self-adaptive architecture for web service monitoring which tries to predict when the provider may not respond in time because of its load. It focuses initially on one non-functional feature, the response time. The goal is to decrease the average response time of requests made to the provider. To improve the response time, the proposed architecture uses a fuzzy method to increase the priority of services whose contract may be violated. Changing priorities gives to the provider an adaptive behaviour to avoid contract violations. Results show that a decrease in the average response time implies a decrease in the number of contract violations.