Hybrid Start（混合慢启动）

    Standard slow start does not work well under large bandwidth-delay product (BDP) networks. We find two causes of this problemin existing three popular operating systems, Linux, FreeBSD and Windows XP. The first cause is that because of the exponential increase of cwnd during standard slow start, heavy packet losses occur. Recovering from heavy packet losses puts extremely high load on end systems which renders the end systems completely unresponsive for a long time, resulting in a long blackout period of no transmission. This problem commonly occurs with the three operating systems.The second cause is that some of proprietary protocol optimizations applied for slow start by these operating systems to relieve the system load happen to slowdown the loss recovery followed by slow start. To remedy this problem, we propose a new slow start algorithm, calledHybrid Start (HyStart) that finds a safe exit point of slow start at which slow start can finish and safely move to congestion voidance without causing any heavy packet losses. HyStart uses ACK trains and RTT delay samples to detect whether (1) the forward path is congestedor (2) the current size of congestion window has reached the available capacity of the forward path. HyStart is a plug-in to the TCP sender and does not require any change in TCP receivers. We implemented HyStart for TCP-NewReno andTCP-SACK in Linux and compare its performance with different slow start schemeswith the TCP receivers of the three different operating systems in the Internetand also in the lab testbeds. Our results indicate that HyStart works consistently well under diverse network environments including asymmetric links and high andlow BDP networks. Especially with different operating system receivers (Windows XPand FreeBSD), HyStart improves the start-up throughput of TCP more than 2 to 3times.

1、混合慢启动设计的目的：

    标准的慢启动在BDP网络环境下表现不好，不好的原因主要有两个：

    1） 标准慢启动的拥塞窗口指数式的增长方式过于激进容易导致大量丢包，丢包恢复性能损耗太大。

    2） 被优化过的慢启动机制，丢包时在数据包重传恢复的时候碰巧试图去减小服务器的负载，导致数据包恢复慢。

    总结这些原因都是因为慢启动过程过于盲目，不能及时的预测拥塞，导致了大量丢包，所以混合慢启动机制的主要

功用是在慢启动阶段试图找到“合理 ”的 退出慢启动 进入拥塞避免状态点（safe exit point）。 

2、慢启动是怎么找到safe exit point

     1）通过ACK train

     2）通过RTT delay 

3、混合满慢启动目前在cubic拥塞控制算法中引入，后面会具体分析cubic混合慢启动实现的细节。