Going to 50-bit
addresses with ten
bits per field?
That makes no sense
from a computing efficiency
viewpoint, as those
are not powers of 2.
48 bits would be on
octet boundaries, but harder to
process than 64 bits
in a word (ATM had trouble with
48-byte
payloads, too. A fudge between those who
wanted 32
bytes and those who wanted 64.)
IP protocol
version 5 is already defined and in use
for RFC1190,
and that use would have to be deprecated
first. Picking
7 is easier; for a long time someone
on this list
was pushing an idea of IPv8... and IPv9.
(Try this: sitting
through talks about how awesome IPv6 is
and how it solves
addressing problems and routing table
problems,
then asking innocently 'why is there a
jump from IPv4 to
IPv6? What happened to 5?' If they can't
tell you that, then
just maybe they shouldn't be trying
to sell you
stuff.)
L.
'don't
understand it fully' may be somewhat understated.
Den 10. des. 2015 18:53, skrev Alexey Eromenko:
> Good news: After much jumping through hoops I was
able to decipher
> what-it-takes to get my draft spec submitted (I still
don't understand
> it fully, though).
Where's the transition plan?