用于分层移动IPv6网络拥塞控制的新机制

针对在IPv6无线网络切换过程中, 短暂的链路连接断开会导致反复丢包; 同时在切换过程中, 由于带宽的改变, 在新的接入点就有可能发生丢包或资源浪费以及拥塞。提出一种以TCP协议为基础的路径损耗确认（TCP-PLACK）机制来代替TCP选择确认机制（TCP-SACK）。每当一个TCP接收方在断开或切换后而连接到一个新的接入点时, 上述的TCP-PLACK机制就会发送一个特殊的确认, 其中包含有在新接入点丢包的详细信息和可用带宽。收到这个确认, 发送方会重新发送丢失的包, 并在新接入点根据带宽的可用性调整发送速率。实验结果表明, 该机制有利于改善IPv6网络切换过程中的丢失恢复和速率控制问题。

New mechanism applied for congestion control in hierarchical mobile IPv6 networks

The packet losses occur repeatedly due to temporal link disconnection during handoff in IPv6 wireless networks. Also during the handoff, due to this bandwidth change, packet drops or wastage of resources and congestion may occur at the new access point. This paper proposed a TCP(transport control protocol) based path loss acknowledgement(TCP-PLACK) mechanism in place of TCP-SACK(selective acknowledgement) mechanism. The proposed TCP-PLACK mechanism sent a special acknowledgment which consisted of packet loss details and available bandwidth at the new access point whenever a TCP receiver was attached to a new access point after a disconnection period or handoff. On receiving this acknowledgment, the sender retransmited the lost packets and adjusted the sending rate according to the bandwidth availability at the new access point. Results of experiments show that the proposed TCP-PLACK mechanism is beneficial for solving problems of packet losses and sending rate adjustment during the switch in wireless networks.