Subject: Re: This PMTU thread
To: NetBSD Networking Technical Discussion List <tech-net@netbsd.org>
From: Lucio de Re <lucio@proxima.alt.za>
List: tech-net
Date: 11/24/1998 19:26:14
I think I should commend Greg on fighting so hard to put a point
across, against all odds :-)
The opposition is trying to make a political issue of a mechanical
thing. My view is that Greg's proposal has merit not as a long-term
solution (black hole detection seems a better approach, and Greg has
made it clear that he would prefer that approach) but as a route around
the evident dilution of expertise brought about by the Internet
explosion.
At the risk of overstepping a mark, I'd like to point out that the very
issue Greg is keen to work around, others are trying desperately to
overlook: too many administrators are not as comfortable with IP as the
experts on this forum would wish them to be. My feelings are that this
situation will get worse before, if ever, it will start improving.
In this context Greg suggestion would be, at the very least, a measure
of the problem: it would permit, minimally, logging and analysing the
occurrence and frequency of such divergence from the norm. I do
appreciate that the NetBSD network developers may have their hands full
tracking the standards, and that this type of divergence may have to
remain in the realm of the experimental, but shouting Greg down seems
to me to be counterproductive.
It would be good, in my opinion, at the very least to meet Greg half
way. If I recall the discussion correctly, Greg found it difficult to
diagnose the problem because the corruption did not cross the boundary
to the destination host; merely making it possible to detect the
problem _before_ discarding the broken data (I must confess that I paid
more attention to the tone of the discussion than the subject matter
itself) would be useful. From there to breaking the IP rules may be a
short, slippery downhill, but I think the first step may be a positive
one.
++L