Jump to content

Maximum segment lifetime

From Wikipedia, the free encyclopedia
This is an old revision of this page, as edited by 138.232.3.38 (talk) at 21:44, 8 February 2023 (correcting the Linux implementation narrative and other fine tuning). The present address (URL) is a permanent link to this revision, which may differ significantly from the current revision.

Maximum segment lifetime is the time a TCP segment can exist in the internetwork system. It was defined in 1981 to be 2 minutes.[1] <quote> For this specification the MSL is taken to be 2 minutes. This is an engineering choice, and may be changed if experience indicates it is desirable to do so. </quote>

The Maximum Segment Lifetime value is used to determine the TIME_WAIT interval (2 * MSL), the minimum time a system must wait before designating the socket closed, thus preventing the socket from being re-used before that interval.

Values in various operating systems

The command that can be used on Solaris systems (prior to v11) to determine the TIME_WAIT interval is:

   ndd -get /dev/tcp tcp_time_wait_interval

60000 (60 seconds) is a common value.

On FreeBSD systems this description and value can be checked by the command sysctl:[2]

   sysctl -d net.inet.tcp.msl
   sysctl net.inet.tcp.msl

which gets the result:

   net.inet.tcp.msl: Maximum segment lifetime
   net.inet.tcp.msl: 30000

In Linux, the kernel will wait up until TCP_TIMEWAIT_LEN before closing a socket in the TIME_WAIT state. This value is hardcoded at 60 seconds, which implies a MSL of 30 seconds. Linux implements several possible optimizations to shorten the TIME_WAIT state through _recycling_, down to a minimum of 3.5s, for an effective MSL of 1 second.[3]

References

  1. ^ "RFC 793". Transmission Control Protocol. Retrieved December 5, 2006.
  2. ^ "Tuning FreeBSD to serve 100-200 thousands of connections".
  3. ^ "Linux Kernel socket implementation".