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

[Alexander Zimmermann <[e-mail]>]

Hi Fernando,

thank you for clarification. I've got it.

Alex

Am 26.03.2010 um 13:31 schrieb Fernando Gont:

> Hi, Alexander,
> 
>> first of all, I didn't read your draft, so please apologizes my
>> maybe stupid remark.
>> 
>> I guess your algo. is not compatible with the timestamp generation
>> in RFC 4987 (3.6.  SYN Cookies). Right?
> 
> Actually, there's an interesting part to it:
> 
> What matters is that the client-side (i.e., the TCP performing the
> active open) generates timestamps such that they are
> monotonically-increasing across connections, as *those* are the
> timestamps that are used for "recycling" connections that are in the
> TIME_WAIT state.
> 
> The timestamp generation implemented by FreeBSD makes sense for the
> server-side (the side implementing the extended SYN-cookies). Therefore,
> both approaches can co-exist: i.e., clients could generate timestamps
> such that they are monotonically-increasing (and thus allow the
> recycling of connections in the TIME_WAIT state), while servers can
> generate timestamps with the FreeBSD scheme, such that they allow for
> extended SYN cookies.
> 
> Does this make sense?
> 
> P.S.: Yes, clearly, this would be worth noting... and in particular,
> could be taken as a motivation for splitting the generation of
> timestamps into a separate I-D.
> 
> Thanks so much for your comments!
> 
> Kind regards,
> -- 
> Fernando Gont
> e-mail: [e-mail] || [e-mail]
> PGP Fingerprint: 7809 84F5 322E 45C7 F1C9 3945 96EE A9EF D076 FFF1
> 
> 
> 
>