A hybrid rate control mechanism for forwarding and congestion control in named data network

• 第一著者: Alsamman, Mohammed Gamal
• フォーマット: 学位論文
• 言語: 英語; 英語; 英語
• 出版事項: 2020
• 主題: TK7885-7895 Computer engineering. Computer hardware
• オンライン･アクセス: https://etd.uum.edu.my/8852/1/depositpermission_s95138.pdf; https://etd.uum.edu.my/8852/2/s95138_01.pdf; https://etd.uum.edu.my/8852/3/s95138_references.docx

Named Data Networking (NDN) is an emerging Internet architecture that employs a pull-based, in-path caching, hop-by-hop, and multi-path transport architecture. Therefore, transport algorithms which use conventional paradigms would not work correctly in the NDN environment, since the content source location frequently changes. These changes raise forwarding and congestion control problems, and they directly affect the link utilization, fairness, and stability of the network. This study proposes a Hybrid Rate Control Mechanism (HRCM) to control the forwarding rate and link congestion to enhance network scalability, stability, and fairness performance. HRCM consists of three schemes namely Shaping Deficit Weight Round Robin (SDWRR), Queue-delay Parallel Multipath (QPM), and Explicit Control Agile-based conservative window adaptation (EC-Agile). The SDWRR scheme is scheduling different flows in router interfaces by fairly detecting and notifying the link congestion. The QPM scheme has been designed to forward Interest packets to all available paths that utilize idle bandwidths. The EC-Agile scheme controls forwarding rates by examining each packet received. The proposed HRCM was evaluated by comparing it with two different mechanisms, namely Practical Congestion Control (PCON) and Hop-by-hop Interest Shaping (HIS) through ndnSIM simulation. The findings show that HRCM enhances the forwarding rate and fairness. HRCM outperforms HIS and PCON in terms of throughput by 75%, delay 20%, queue length 55%, link utilization 41%, fairness 20%, and download time 20%. The proposed HRCM contributes to providing an enhanced forwarding rate and fairness in NDN with different types of traffic flow. Thus, the SDWRR, QPM, and EC-Agile schemes can be used in monitoring, controlling, and managing congestion and forwarding for the Internet of the future.