| As a hot topic of information centric networking(ICN), named data networking(NDN) has attracted increasing attention of research institutes and researchers all over the world. Congestion control is one of the critical issues not only in TCP/IP networking but also in named data networking. Congestion control used in TCP/IP networking cannot be applied in NDN without any modification because the architecture of NDN is different from the architecture of TCP/IP networking. We should consider the feature of NDN while designing congestion control strategy for NDN. Moreover, an efficient forwarding strategy can alleviate network congestion. The contributions of this thesis are mainly twofold.In the first part of thesis we propose a rate control strategy to alleviate congestion in named data networking. According to the feature of forwarding in NDN, hop-by-hop congestion control strategy can quickly react to network congestion. By carefully selecting utility function as objective function, we model the single node's resource allocation problem as a network utility maximization problem. Then we propose the hop-by-hop interest-shaping based congestion control strategy via solving the network utility maximization problem. The simulation results show that the proposed strategy can effectively mitigate packet loss in NDN.The second part of thesis focuses on mitigating network congestion in NDN via adaptive forwarding. Dynamically avoiding congestion link while forwarding can effectively alleviate or even avoid congestion. Packet delivery in TCP/IP networking sticks to the routing table which is built in advance; however, NDN's routers are aware of the changing in environment and can adaptively forward packets by gathering information from their neighbors. By considering the adaptive forwarding in NDN, we present the Q-value based Interest forwarding strategy in order to minimum the delivery time of data packet. Simulation results show that our strategy has lower average delay and packet loss. |
PDF Full Text Request