[99346] in North American Network Operators' Group
Re: Question on Loosely Synchronized Router Clocks
daemon@ATHENA.MIT.EDU (Robert Bonomi)
Tue Sep 18 18:33:19 2007
Date: Tue, 18 Sep 2007 17:32:03 -0500 (CDT)
From: Robert Bonomi <bonomi@mail.r-bonomi.com>
To: nanog@merit.edu
Errors-To: owner-nanog@merit.edu
> From owner-nanog@merit.edu Tue Sep 18 10:57:15 2007
> Date: Tue, 18 Sep 2007 08:55:19 -0700
> From: "Xin Liu" <smilerliu@gmail.com>
> To: "Bora Akyol" <bora.akyol@aprius.com>
> Subject: Re: Question on Loosely Synchronized Router Clocks
> Cc: nanog@merit.edu
>
>
> Ideally, yes, a protocol should not rely on clock synchronization at
> all. However, to ensure freshness of messages, we don't have many
> choices, and clock synchronization seems to be the least painful one.
> So we asked about router clocks on the current Internet. If normally
> router clocks are synchronized and we have a mechanism to detect and
> fix out-of-sync clocks, is it reasonable to assume clock
> synchronization in the rest of our design?
You are free to "assume" anything you feel like in the design of a new
protocol.
The greater the divergennce between your 'assumptions' and *UNIVERSALLY*
IMPLEMENTED conditions in the real world, the more barriers there are to
acceptance and deployment.
Within a single administrative domain, routers are 'usually' -- but *NOT*
"almost always" -- moderately closely synchronized. Across different
administrative domainns, any such synchronization is 'happy accident',
nothing more.
As far as 'assuming clock synchronization' goes, one of the other subscribers
to this list has a _very_ applicable remark: "I encourage my competitors to
design like this." <grin>
