Handling feature interactions in the SIP servlet context

• Main Author: Yang, Lu
• Format: Others
• Published: 2005
• Online Access: http://spectrum.library.concordia.ca/8560/1/MR10257.pdf; Yang, Lu <http://spectrum.library.concordia.ca/view/creators/Yang=3ALu=3A=3A.html> (2005) Handling feature interactions in the SIP servlet context. Masters thesis, Concordia University.

Session Initiation Protocol (SIP) and Servlet technology introduce new ways of delivering telephony services over IP networks. The richness and flexibility of these new protocols make it faster and easier for service providers to develop and deploy new services. This flexibility, however, is offset by the challenge of managing the feature interaction problem, which can prove to be quite severe. This thesis proposes a modified SIP Servlet architecture and introduces a logical entity, the Feature Interaction Handler (FIH), to address the feature interaction problem for telephony services. The approach addresses offline and online feature interaction detection, the former occurring when the user registers to a feature and the latter occurring during feature runtime execution. For offline feature interaction detection, a behaviour mapping approach is introduced to reduce the interaction matrix table. For online feature interaction detection, two mechanisms are proposed---'forward detection' and 'backward detection'. Forward detection extends the originating side user profile when sending the message, such that the terminating side can use it for detection. In contrast, backward detection correlates the SIP session 'request' message with the 'response' message belonging to the same session in order to determine if the resulting service behaviour is acceptable. To validate the new feature interaction detection approach, an offline feature interaction detection tool and online feature interaction detection unit FIH have been implemented. The feature interaction benchmark is applied on both the tool and the FIH, the result proves to be successful. The feature interaction detection approach proposed in this thesis proves to be a viable solution in the context of SIP servlet service environment.