QoS Level Estimation for Web Service Composition in an Environmental Assessment System

In the Intelligent Distributed Environmental Assessment (IDEA) system, sensors are used to measure noise and air pollution. The architecture is service-oriented, allowing application architects to combine services provided by the system platform in order to satisfy application needs. A typical user application may require a set of those services. Because of the sensor nodes' resource limitation, some services may not keep Quality of Service (QoS) levels stable over a long period, which could lead to Service Layer Agreement (SLA) violations. To maintain SLA, QoS levels must be estimated so that services can be properly invoked and combined to satisfy the user's QoS requirements. This paper proposes a novel approach to estimate QoS levels of web services. On the basis of monitored values of QoS parameters, a Naïve Bayes estimation model is developed to select the best-suited services. This paper also presents approaches adopted to monitor QoS levels of web services in the IDEA environment.