W3C home > Mailing lists > Public > ietf-http-wg-old@w3.org > September to December 1994

Re: TIME_WAIT

From: Chuck Shotton <cshotton@oac.hsc.uth.tmc.edu>
Date: Tue, 20 Dec 1994 19:52:09 -0600
Message-Id: <ab1d3a32040210041805@[129.106.201.2]>
To: Jeffrey Mogul <mogul@pa.dec.com>, Henrik Frystyk Nielsen <frystyk@www0.cern.ch>
Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com
At 6:18 PM 12/20/94, Jeffrey Mogul wrote:
>    I still think that it is better to keep the connection open for a small
>    amount of time as the server socket will go into TIME_WAIT state when
>    it is closed and then can't be reused before about 240 seconds
>    (recommended timeout accorsing to Simon)
>
>This not just Simon's recommendation.  This is a MUST (i.e., mandatory)
>requirement of RFC1122 (section 4.2.2.13).
>
>And, to clarify: what cannot be reused is the memory allocated to the
>connection state record; all other resources are reusable.  However, in
>most BSD-based systems, this memory usage can be significant (one
>mbuf).

Just a reminder. This issue only really applies to Unix implementations,
and not O/S2, VMS, Windows, Windows NT, Mac O/S, VM, or any other non-Unix
server and/or TCP/IP implementations not descended from the BSD network
kernel.

-----------------------------------------------------------------------
Chuck Shotton
cshotton@oac.hsc.uth.tmc.edu                           "I am NOT here."
Received on Tuesday, 20 December 1994 17:51:18 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 24 September 2003 06:31:10 EDT