[14160] in North American Network Operators' Group
Re: caching at windmills (i mean exchange points)
daemon@ATHENA.MIT.EDU (Ton Verschuren)
Wed Dec 3 12:10:08 1997
To: Randy Bush <randy@psg.com>
cc: k claffy <kc@nlanr.net>, nanog@merit.edu, ircache@nlanr.net
In-reply-to: Your message of "Tue, 02 Dec 1997 09:14:00 PST." <m0xcvu3-0007zbC@rip.psg.com>
Date: Wed, 03 Dec 1997 09:58:11 +0100
From: Ton Verschuren <Ton.Verschuren@surfnet.nl>
==> From: Randy Bush
| > I read your proposal an have a question. I assume that your MAE-west
| > cache will act as a parent for the MW ISP's (otherwise, there's no data
| > in the cache). That means that all outbound traffic from MW is routed
| > through your cache, isn't it? That also means that e.g. all
| > transatlantic traffic will flow through your lines. Or am I missing
| > something?
|
| You are not differentiating between an proxy cache and an interceptor
| cache. The NLANR cache is a proxy cache. Clients/peers/children must
| consciously point at it.
I know. Still, for a cache to fill it has to fetch docs and thus act as
a parent. If the NLANR cache acts as a parent for all MW connected
ISP's than *all* misses from those ISP's get routed through the NLANR
cache and for the transatlantic traffic the lines of NLANR's ISP will
fill up. No?
Ton.
