A centralized resource and admission control scheme for NGN core networks

Recently, telecommunications network is considered to be evolving to the Next Generation Network(NGN) which capable of converging heterogeneous networks and provides converged services. The leading telecommunication operators have already tried to construct an NGN. It is one of the trends that DiffServ-aware MPLS is adopted as an NGN core network technology. Resource and Admission Control Function(RACF) is needed to support QoS of the SIP based converged services, which are per session based real time services, such as IP telephony and video telephony. Many studies of the resource control schemes in Internet architecture have worked so far, but the study of the QoS control scheme appropriate NGN appears nowadays and has a strong tendency of focusing on the edge and the access networks. In this paper, we propose a centralized QoS control scheme for the purpose of effective traffic engineering in NGN core networks, which refer to Centralized MPLS-Traffic Engineering (CMPLS-TE). We also propose a hierarchical QoS control architecture for improvement of simplicity and scalability in the whole view of the NGN using a divide and conquer strategy which separates interesting objects that are the core and the access networks. We present the modelling of the per session based centralized control scheme suggested in this paper and the distributed traffic engineering scheme(e.g. RSVP-TE, CR-LDP) for the verification of control costs. We also present the design prototype of CMPLS-TE model based on a simulator, as we expect the study on the performance evaluation of per session based centralized QoS control scheme will arise in the near future.