OK, so IPv4 is probably right to be acting like the old man in Monty Python and the Holy Grail, and screaming “I’m not dead yet!”, but we certainly shouldn’t hold out any hope that it’ll be getting any better. Clonk it on the head as soon as possible, because really, it’s been extremely poorly for many years now.
I’ve mentioned before that my biggest argument that IPv4 has already exhausted itself is the mere presence of aggregating NATs – Network Address Translators, whose sole purpose is to take multiple hosts inside a network, and expose them to the outside as if they were really only processes on one host with one IP address. If IPv4 were large enough, we wouldn’t have needed these at all, and at best, they were a stop-gap measure, and an inconvenient one at that.
Well, now we can’t really stop the gap any longer. We’ve hit the first of a set of dominoes that leads to us not even having enough IPv4 addresses to support the Internet with NATs in place.
That’s right, no more /8 networks are left in IANA’s pool to assign. OK, I know it says ”5/256” are left, but that’s only because the IANA (Internet Assigned Numbers Authority) haven’t yet announced that they’ve given out those last five, and they have previously announced that when they get down to five, those five will automatically be distributed.
Yesterday, the counter said “7/256”, but earlier today, APNIC – the RIR for the Asia & Pacific region (RIR – Regional Internet Registry) – bought two entries to serve their ever-growing Internet market. That will trigger the IANA to distribute the remaining five blocks.
And no, Egypt’s IP blocks are not available for re-use.
Yes, that’s right, this isn’t a “shut everything off and go home” moment – as I said before, this is merely the tipping of an early domino in a chain. Next, the last five /8s will be given to the five RIRs, and then they will use those to continue handing out addresses to their ISPs. At some point, the supply will dry up, and it will either become impossible, or expensive, to get new public-facing addresses. Existing addresses will still work even then, of course, and several of the new IPv4 address assignments will, ironically, be aggregating NATs that will allow the IPv6 Internet to access old IPv4 sites!
The only question now is what we call this momentous slide into IPv4 exhaustion. Certainly, RagNATok has a pleasant ring to it, as it invokes the idea of a twilight of the old order, a decay into darkness, but this time with a renewal phase, as the new Internet, based entirely on IPv6, rises, if not Phoenix-like from the ashes, then at least alongside, and eventually much larger than the old IPv4 Internet.
As you can tell from my tone, I don’t think it’s doom and gloom – I’m quite looking forward to having the Internet back the way I remember it – with every host a full-class node on the network. It’s going to mean some challenges, particularly in the world of online security, where there will be new devices to buy (bigger addresses mean larger rule-sets, and existing devices are already pretty much operating at capacity), new terminology to learn, and new reasons to insist on best practices (authentication by IP address was never reliable, and is particularly a bad idea when every host has multiple addresses by default, and by design will change its source address on a regular basis).
Perhaps the Mayans were right in deciding that 2012 is the year when everything changes (to borrow a line from Torchwood).
The rather unassuming name that has been chosen for this particular date – when the last assignment leaves the IANA – is “X-Day” – X as in “eXhaustion”.
The next date for your calendars, then, is World IPv6 Day, two days after my birthday, or for those of you that don’t know me, June 8, 2011, which is when major Internet presences including Google, Yahoo and others, will be switching on full IPv6 service on their main sites, and seeing what breaks. Look forward to that, and in the meantime, test some known IPv6 sites, like http://ipv6.google.com to ensure that you’re getting good name resolution and connectivity.
If you’re running an FTP server on Windows, I encourage you to contact me at email@example.com if you would like to test WFTPD or WFTPD Pro for IPv6 connectivity. We are currently beta-testing a version with much greater IPv6 support than before.