Abstract

New network architectures, such as the Intelligent Network (IN), have evolved in response to the changing needs and demands for advanced and sophisticated telecommunications services. However, as more services are introduced into the network, a new problem of interactions between various services/features, becomes more prominent. This problem arises when multiple services or features interfere with each other and produce unexpected results, which disturb the users. This thesis presents my work in modeling features and detecting feature interactions using Message Sequence Charts (MSCs). The modeling technique is based on the Advanced Intelligent Network (AIN) architecture and call models. To effectively detect feature interactions, we propose an MSC feature specification style which embodies several important aspects of features directly related to feature interactions. Based on the modeling of features using the specification style, we propose a new approach for detecting feature interactions. This approach includes definitions, classification of feature interactions, and specific detection algorithms for various types of interactions. We developed a prototyped feature interaction detection tool to implement our approach. With this tool, we are able to detect many interactions described in the Bellcore feature interaction benchmark. Our detection technique has maintained its consistency and accuracy in detecting these interactions. However, some limitations of our approach prevent us from detecting certain types of interactions. Combining our feature specification style, detection approach and tool, we propose a general framework for feature specification and interaction detection for IN services.