On Mon, 12 Sep 2005, Brad Knowles wrote:
> Date: Mon, 12 Sep 2005 09:05:51 +0200
> From: Brad Knowles <[EMAIL PROTECTED]>
> To: [email protected]
> Subject: Re: [time] Nearby servers
> Sender: [EMAIL PROTECTED]
>
> At 8:01 AM +0200 2005-09-12, Henk P. Penning wrote:
>
> > Great ; can we assume (for all practical purposes) that
> > the relation is symmetrical ? That is
> >
> > .xx => .yy implies .yy => .xx
>
> I'm not convinced that this is the case. Moreover, we're more
> concerned about being topologically close to a decent number of
> servers, not just being regionally close. Moreover, you have a
> number of cases where countries like Spain are close to both Portugal
> and France, but France and Portugal are not close to each other.
Please read again ; I asked if it is true that
A is close to B -->IMPLIES--> B is close to A
If it is true, the config file can be half as big ; thats all.
> Looking at this list, it seems to me that any zone that doesn't
> have at least five servers in it should probably be demoted to zero,
> and then that domain should become an alias for the closest larger
> domains.
For what countries would 'region close first' be worse than
'europe random' ?
If it is true in the majority of cases, it is a good first
approximation. Exceptions can be easily specified.
> Brad Knowles, <[EMAIL PROTECTED]>
HPP
---------------------------------------------------------------- _
Henk P. Penning, Computer Systems Group R Uithof CGN-A232 _/ \_
Dept of Computer Science, Utrecht University T +31 30 253 4106 / \_/ \
Padualaan 14, 3584CH Utrecht, the Netherlands F +31 30 251 3791 \_/ \_/
http://www.cs.uu.nl/staff/henkp.html M [EMAIL PROTECTED] \_/
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers
