Subject: Re: traceroute max ttl uplift
To: Darren Reed <darrenr@reed.wattle.id.au>
From: Robert Elz <kre@munnari.OZ.AU>
List: tech-net
Date: 05/21/2002 12:10:55
Date: Mon, 20 May 2002 22:07:58 +1000 (EST)
From: Darren Reed <darrenr@reed.wattle.id.au>
Message-ID: <200205201207.WAA23726@avalon.reed.wattle.id.au>
| Well...hmmm...what serves most of the world's web pages ?
Once upon a time, it used to be *BSD of some variety. No idea these days.
| Well, some points of reference:
| Solaris (2.5.1+), HP-UX (11+) both ship with a default TTL of 255;
| Windows 2000 has a default ttl of 128;
| Linux has used 255 since at least 2.2.
Those being broken doesn't mean that NetBSD should be as well.
I understand why a vendor would do that .. it means not having to explain
to the clueless masses why they can't get to some particular random sites
when they can get to others. I think NetBSD can do better than that.
| I'm tending towards using 64 in traceroute and maybe bump the kernel to 128.
For traceroute, fine - or even more than 64 perhaps.
There's no current reason to increase the kernel limit - 64 is still way
more than the diameter of the worst parts of the net (which probably runs
to about 40, and increases very very slowly).
kre