The IP over ATM Mailing List Archive by date

[Grenville Armitage <gja@thumper.bellcore.com>]

Re-ordering a bit:

>>I think not.  Let us just burn the extra bytes and get it over with.  It is
>>only the control plane after all.  I like the way the data plane works
>>out but I don't think that saving the bytes in the control plane justifies
>>the increase in complexity.

Okay. I'm in the process of summarising the last few weeks discussions,
so I'll produce some samples of the various encoding schemes too.
That way we'll have something visual to look at. I should have this
done by the end of today and will put it on my web site.

( As an aside:

	[..]
>>However I do not think it is acceptable for the control plane.  Recall that
>>the sequence number is protocol-specific.  Now I need to look at the ar$pro
>>and if it is 0x80 go look for a TLV and _then_ do sequence number processing?

In one of my emails I suggested the SNAP TLV as always being the first TLV
in the list whenever ar$pro = 0x80. In this case the packet processing
is only marginally different, since the SNAP is always immediately
referenced by indirecting through ar$extoff. No list scanning needed,
but point noted that its one extra mental hoop for the implementor
to think about. )

cheers,
gja