My own answer is “I don’t know”. I only know that there were a few of us thinking about the problem in 1989. Roy Smith raised the issue on the TCP-IP mailing list on November 25th of that year with this message:
Date: 25 Nov 88 14:56:57 GMT From: roy@phri.UUCP (Roy Smith) To: comp.protocols.tcp-ip Subject: Running out of Internet addresses?
Has anybody made any serious estimates of how long it will be
before we run out of 32-bit IP addresses? (Silly question; I'm sure a very
great amount of thought has been given to it by many people.) With the
proliferation of such things as diskless workstations, each of which has
its own IP address (not to mention terminal multiplexors which eat up one
IP address per tty line!), it seems like it won't be too long before we
just plain run out of addresses.
Yes, I know that 2^32 is a hell of a big number, but it seems like
we won't get anywhere near that number of assigned addresses before we
effectively run out because most nets are sparsely populated. My little
bit of wire, for example, has 256 allocated addresses yet I'm only actually
using 30 or so.
--
Roy Smith, System Administrator
Public Health Research Institute
{allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy@uunet.uu.net
"The connector is the network"
Back then we used IP addresses in a considerably sparser way than we do today. That message kicked off a lengthy discussion in which nobody seriously was in denial about the potential for a problem. You can find the whole archive of the exchange here. There were two concepts that were touched upon. The first was whether or not we could use the so-called “Class E” space (240.0.0.0/4). I and others gave this serious thought at the time. However, the related issue which won the day was that fixed address lengths were an important property to be maintained. Vint Cerf raised that design consideration as a question. He also raised the possibility of using variable-length OSI addresses.
Eliot –
If memory serves, there were other, bigger fish to fry along the way – including congestion collapse, routing table explosion, and how to keep up with demand for bandwidth. All of those problems were going to have to be solved more or less independent of address space size – and it’s likely that you really couldn’t have pulled off a bigger address space size until you got routing worked out in IPv4.
Small world 🙂 I worked with Roy at Micromuse and IBM in the early 2000s.
Wow, what a blast from the past. Fuat, thanks for sending this to me.
Fast forward a quarter of a century to find me working at EMC adding IPv6 support to their network management product (nee Smarts InCharge). Then we discovered that none of our customers actually cared about IPv6 so there was no market. I’m not sure they sold a single copy.
I also had to fight with our own internal IT department to get IPv6 enabled (and configured properly) on the development network so we could use it internally. In 2010 when I left EMC, IT was still insisting on using 3FFE (6bone) addresses internally. And IPv6 traffic was strictly isolated to what IT called “the lab network” because they were afraid of the security implications of allowing it on “the corporate network” (i.e. their packet inspecting firewalls couldn’t monitor IPv6).
Roy, you got me started in academic computing in the very early 90s. Don’t know if you remember, but seeing this is like reviewing old Kermit code and finding your comments in the margins. That happened in the late 90s. Nice to see you pop up from time to time.