The missing link that both of these things can be attributed to is the reduction in TCP timeout-based recoveries on the newer kernel because of the RFC 3517-related changes.
这些情况的原因中缺少的一环是在新内核中减少了TCP 基于超时的恢复机制，原因是与RFC3517有关的改变。
If configured, the RTO value would be set as the initial retransmission timeout for all the TCP connections on the system.
如果配置了该值，那么会将这个rto值设置为该系统中所有TCP连接的初始重新传输超时。
This routine walks through the table of TCP control blocks and, for each connection, it decrements the timeout, specified in ticks, by one.
这个例程将遍历tcp控制块表，并且对每个连接的超时(以计时单元为单位进行指定)进行递减操作，即每次减一。
As you know, TCP dynamically calculates timeout based on the round-trip time measured by TCP.
正如您所知道的，TCP根据所测量的往返时间，动态地计算超时。
Currently, TCP retransmission timeout takes on a minimum value of three seconds.
目前，TCP重新传输超时的最小时间为3秒。
Slide window, timeout retransmission and acknowledgement mechanism are used to make sure of the credibility of TCP.
我们采用了滑动窗口、超时重发和确认机制确保了TCP协议的可靠性。
Slide window, timeout retransmission and acknowledgement mechanism are used to make sure of the credibility of TCP.
我们采用了滑动窗口、超时重发和确认机制确保了TCP协议的可靠性。
应用推荐