[782] in Commercialization & Privatization of the Internet
Re: IETF questions -- Internet growth
daemon@ATHENA.MIT.EDU (Fred Baker)
Thu May 30 13:09:51 1991
Date: Thu, 30 May 91 10:05:52 PDT
From: fbaker@acc.com (Fred Baker)
To: bzs@world.std.com, francis%zaphod@gargoyle.uchicago.edu
Cc: com-priv@uu.psi.com, emv@ox.com, ietf@ISI.EDU,
Barry:
it's actually 2,097,150 class C nets...
Brian writes:
>> using the scheme of arbitrary length network numbers I would opt for a
>> 20 bit network number, a six bit subnet, and six bits for host number.
Lessee here, seems to me that those "C" nets could be recovered. Right
now, according to RFC 1166, Class C nets are now allocated from
192.x.x.x through 195.x.x.x. A LOT of these, 192.70.0.rrr -
195.5.67.rrr, are unassigned holes.
If we concede that, it seems that we could redistibute 196.x.x.x
through 223.x.x.x as a "Class F net", which would have a 12 bit host
part. That would make the distribution be:
Class A addresses 01000000->7e000000 126
Class B addresses 80010000->bffe0000 16,382
Class C addresses c0000100->c3fffe00 262,142
Class F addresses c4000100->dfffe000 114,688
If the obvious subnet approach is used - 14 nets with mask
255.255.255.0, its unlikely that installed hosts or routers would be
able to detect the difference, and new router releases could capitalize
on the new format for table size reduction.
If the real issue is that the world at large is competing for Class B
addresses, mayhap folks would be willing to settle for Class F?
Fred
