| As a concrete realization of NGN, IP Multimedia Subsystem (IMS) can realize end to end user connection, and become the basis for network integration. IMS is based on SIP(Session Initiation Protocal), uses layered architecture, and provides standardized interfaces, which determines the value-added services to be diverse, personalized, and the terminals to be intelligent, the feature interaction problem becomes even more serious, as a result, the study of feature interaction problem among value-added services become an important research topic.In this dissertation, based on the background of services in IMS and the service provision principle, we study and design the IMS feature interaction detection system for detecting feature interaction that has occurred and to be occurred in time, so that every service can operate normally in the call process. In order to facilitate detection, we describe the service implementation process and interaction rule formally. Through study of IMS and feature interaction problem theory, we devide the feature interaction among IMS mult-user call control services to two categories: link interaction and logic interaction, and the link interaction also includes forwarding loop interaction and self-connecting interaction. According to the classification of interactions, we develop the pre-judgment rule. Through analysis and research large number of feature interaction examples deeply, we propose an experiential learning based feature interacion detection approach. Compared with traditional methods, this approach can detect feature interaction earlier. The overall frame of detection approach was designed to grant the concrete functional demand of the approach, of which each functional module has been expatiated on detail. For the feature interaction that has occurred, we propose the detection method, learn the information of interaction, and abstract the interaction rule, which will be send to the interaction set. For the feature interaction that to be occurred, mainly link interaction and the interaction that has occurred and learned, we use the pre-judging method to detect the interaction. In order to verify the validity and feasibility of this approach, we conduct the case study. At last, we designed and implemented the IMS feature interaction detection prototype system, validated the theory and the method mentioned in the paper. |
