> First is, the change of MTU was not one of 576 to 1280.

I think he was talking about changing the minMTU of Steve Deering's Simple IP 
from 576 first into the mistaken 1500 and then into 1280 in the process of 
turning it into IPv6.
http://tools.ietf.org/html/draft-deering-sip-00
(BTW, is that process documented anywhere? RFC1710 doesn't seem to mention the 
MTU.)

> The other point is, I don't see how transparent adaptation layers are an 
> issue at all? I don't think anyone is saying that it's impossible to transmit 
> IPv6 over ATM cells? Or are they?

The problem comes up where the transparency is a leaky abstraction 
(http://en.wikipedia.org/wiki/Leaky_abstraction).

For ATM, the abstraction was not very leaky as the performance impact of 
sending an MTU-size packet was negligible (at least for the more recent cases 
of ATM).

For the constrained networks Ran and I are interested in, the performance 
impact of adaptation layer fragmentation can be significant, and application 
layer protocols that get to choose their packet sizes can benefit quite a bit 
if they know this fragmentation is taking place and what packet sizes are the 
thresholds.

Please read the short ALFI draft for more background.
http://tools.ietf.org/html/draft-bormann-intarea-alfi-03
http://www.ietf.org/proceedings/85/slides/slides-85-intarea-1.pdf

Grüße, Carsten

--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------

Reply via email to