Research On Lightweight Web Services Composition In Internet Of Things

Each device provide its function in a form of web service in IoT based on SOA. Services can be discovered, selected,composed to realize specific fuction just like web services in internet. So, It is vital for integrating magnanimity informations in IoT.Web service is the mainstream method of SOA, it is an effective way to compose magnanimity components, many researchers propose to apply web service standard to devices in IoT. But there are many differences between device services and web services, for example the following aspects:1) device service is related to physical world but web service derive from virtual entity; 2) device service is dynamic exposing to the open environment but web service remain static state; 3) device service is deployed in resources limited devices. Service resources competition for service requests is unavoidable to happen because of the previouse differences, let alone to user’s personalization. Therefore, the traditional WS-* is complex and has a high resources consumption. so the traditional web services compositon methods can’t meet the demands of device services compositon in Io T for the reason that traditional web services compositon methods don’t take these facts into consideration.To satisfy the quality of service and user’s personalization in IoT, This paper seek a model for device service integrating in IoT for the purpose of providing a high qulity of service and satisfying user’s preference when users request device services. The main work is as following:1) We extended the architecture of web service by adding service request proxy and service scheduling proxy. 2) We established QoS model for device service and optimal web service selection method based on combination weighting method. 3) We extended weighted round robin(WRR) algorithm to establish device service scheduling model.4) We established resources optimizating-oriented device service compositon model based on the extended web service architecture and service scheduling model.5) we analyzed the delay performance of our model by using the queue theory. The simulation experiments illustrated that the method is efficient. It not only can guarantee the fair attainability of requesters, but also relieve the delay and delay jitter performance. Meanwhile, to some extent, it can meet the individual requirements in the case of limited service resources.