Intelligent Policy Resource Management for Advanced Broadband Access Networks

This paper proposes an intelligent resource allocation and bandwidth management scheme for use in advanced broadband access networks. One key challenge facing current broadband access service providers is the need to deal with peer-to-peer applications (e.g., Skype and Bit Torrent) and other popular but bandwidth-intensive applications such as Sling Media. Currently, peer-to-peer traffic occupies over 80% of backbone Internet traffic. However, the star topology that is common in many broadband access deployments creates a local bottleneck at the central headend when handling such applications. We first examine how resources can be scheduled for different traffic types in a broadband access network. Based on these studies, we then propose an intelligent resource management (IRM) algorithm. The intelligent algorithm can also be used to provide different service tiers and prevent serious security attacks. We employ OPNET simulation and an experimental implementation to evaluate the merits of the proposed scheme. It is shown that the IRM scheme performs better than conventional scheduling schemes. Specifically, the scheme can manage and control bandwidth-intensive applications, and provide fair bandwidth allocation. The broader impacts of the IRM scheme are the ability to mitigate serious security problems (e.g., worm viruses and denial of service attacks), and the ability to deliver quadruple play services while meeting required performance metrics. This capability will become increasingly important with the emergence of pervasive mobile broadband access networks such as Wi-Max, IMS, and meshed Wi-Fi.