Re: [tcpm] poll for adoption of draft-gont-tcpm-tcp-timestamps-03

[Fernando Gont <[e-mail]>]

Hello, Christian,

Christian Huitema wrote:
>> Clearly, for the sake of this policy, it's better if clients
>> generate timestamps such that they are monotonically-increasing
>> across connections. If the chances of TIME_WAIT tossing are
>> reduced.
> 
> What are the side effects if a computer reboots, keeps the same IP
> address, looses memory of the previous time stamps, and then tries to
> reestablish the previous connection?

As per RFC 793, hosts must not establish new connections for a period of
2MSL after they boot (this is the "quiet time" concept defined in
RFC793). As a result, specs-wise, that scenario should never occur.

If a host does not comply with the "quiet time concept" (i.e.,
reality-based world), then:

* If the selected timestamp for the new connection is
monotonically-increasing wrt the last timestamp seen on the previous
incarnation of the connection, the TIME_WAIT state is tossed (i.e., the
"heuristics" kicks in).

* Otherwise, as explained in my previous response to Joe's post, the
connection request may time out or be rejected (i.e., this is no worse
than the current state-of-affairs without this policy).

i.e., I don't see any side-effects....

Thanks!

Kind regards,
-- 
Fernando Gont
e-mail: [e-mail] || [e-mail]
PGP Fingerprint: 7809 84F5 322E 45C7 F1C9 3945 96EE A9EF D076 FFF1