摘要： 针对现有的组播拥塞控制协议计算量大、实现复杂且实际控制效果不好的缺点，提出一种简单高效的显式速率组播拥塞控制协议。该协议只在组播树的发送端和接收端执行，通过计算拥塞吞吐率来监控拥塞、调节发送速率，避免了对所有往返时延和丢包率等复杂参数的计算，各节点只需少量内存来维护必要的数据和状态信息。NS2仿真证明该协议具有良好的TCP公平性和扩展性，可有效跟踪最拥塞链路，解决“反馈内暴”和“drop to zero”等问题。

关键词: 显式速率, 组播拥塞控制, 单速率

Abstract: Most of the present multicast congestion control protocols have more calculations, more complex implementation and less effect. A practical and effective explicit rate multicast congestion control protocol TRAC-Based Explicit Rate Multicast Congestion Control(TBMCC) is proposed. The protocol can be implemented only at the source and the receivers of the multicast tree by computing throughput rate at congestion to monitor congestion and adjust sending rate without computing all the round-trip time and the rates of loss packets. It needs little memory to update the necessary data and state information. Simulations and experiments prove that TBMCC achieves TCP friendliness and expansibility. It can tract the congestion link effectively and can not suffer from the drop-to-zero problem and is very effective with feedback suppression.

Key words: explicit rate, multicast congestion control, single rate

中图分类号: 

• TP393.02