Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: (ASCEND) Pool handling strategies (was: IP Addresses needed for 4048)
At 09:49 AM 4/1/98 -0500, Phillip Vandry wrote:
>One thing I'm not is a TCP expert, so I may very well be wrong, but
>shouldn't the sender TCP react to the ICMP Host Unreachable messages
>it's going to receive in response to anything it tries to send, thus
>terminating the connection just as effectively as an RST would?
No. Host Unreachable is a temporary condition which could be repaired at
any time. It just means there's a break in the path to the host at the
moment. Remember that TCP is intended to survive nuclear war and in such an
event it's not unreasonable that there might be temporary outages in the
network. :-) Host Unreachable is a valid reason to fail at opening a
connection, but it's not a valid reason to close an established one.
Session close should only be done by end to end agreement (or disagreement
e.g. if one party had rebooted and lost the connection state).
I'm told that in the early days of TCP development at BBN a telnet session
happened to be left open between two machines when some piece of the
network between them was taken down for other experiments. A month later
when the two hosts were again connected by a working network the telnet
session was still alive.
++ Ascend Users Mailing List ++
To unsubscribe: send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd: <http://www.nealis.net/ascend/faq>
Follow-Ups:
References: