Re: [TERNLI] Forwarding corrupt packets

[Michael Welzl <[e-mail]>]

Joe,

I couldn't agree more with everything you say here.

Cheers,
Michael


On Mon, 2006-09-04 at 18:01, Joe Touch wrote:
> Michael Welzl wrote:
> >> The question is the impact of the bad packet.
> ...
> >> If the transport reacts to corrupt packets - as would be the case with
> >> UDP-lite and some configs of DCCP - that would be very bad. The
> >> difference is whether we're talking about corrupted data and corrupted
> >> headers; corrupt data isn't too bad (that's what UDP-lite and DCCP are
> >> designed for). Corrupt headers are VERY bad to forward.
> > 
> > I still don't get it - no matter how you use them, the
> > reaction of DCCP and UDP-Lite would always be to silently
> > discard a packet if the transport header is wrong, so
> > no harm is being done (except for the potential waste of
> > bandwidth you mention above, but we agree that this isn't
> > substantial).
> 
> Bad body, good header: the harm is the bandwidth to the correct
> endpoint; the benefit is that this endpoint knows something legitimate,
> and the packet serves as its own corruption signal
> 
> Bad header: the harm is the signal is sent to the WRONG endpoint.
> 
> I reviewed the Feb 2005 discussion on this issue from End2end-interest.
> 
> Here's the summary:
> 
> - IPv6 does NOT, in RFC2640, require link checksums (as UDP-Lite does).
> 
> The relevant text is:
> 
> -------------------------------
> > Although IPv6 (at 
> > least as I recall) omits the IP checksum because of sufficient link 
> > checksums to detect corrupted packets, there's no requirement that link 
> > layers have such checksums (as hinted in the UDP-Lite RFC) - at least 
> > not in 2460 (anyone else know anywhere else a body is buried in this 
> > regard?)
> -------------------------------
> 
> I saw no followup to that question.
> 
> - I said that UDP-Lite's requirement of link checksums was an ugly layer
> violation. I still think that's the case.
> 
> As far as I can tell, UDP-Lite is relaxing a requirement in IPv6 -
> allowing partial link checksums rather than full, but I can't locate
> that requirement, as per above.
> 
> ---
> 
> I don't have a problem with forwarding the information to the wrong
> place; I have a problem with the wrong place acting on it. You're
> assuming that the UDP-Lite partial checksum is sufficient to avoid that;
> for some reason, the UDP-Lite authors and WG clearly did not.
> 
> It'd be useful to review the UDP-lite discussions as to why that text
> was required so prominently.
> 
> Joe
-- 
Michael Welzl <[e-mail]>
University of Innsbruck