Waste will kill ipv6 too

Sat Dec 30 01:25:45 UTC 2017

I'm saying I should be able to use whatever size LAN I want.

Go ahead. Just don't use anybody else's addresses to do it. :)

 -mel

On Dec 29, 2017, at 4:52 PM, John Lightfoot <jlightfoot at gmail.com<mailto:jlightfoot at gmail.com>> wrote:

Excuse the top post, but this seems to be an argument between people who understand big numbers and those who don't.  IPv4 has 2^32 addresses, IPv6 has 2^128, which means 79 octillion people can each have their own internet.  I think Owen is being modest when he says no one alive will be around for the exhaustion of IPv6, I think we're debating whether it will run out in a thousand years or a million.

On 12/29/17, 10:44 AM, "NANOG on behalf of Owen DeLong" <nanog-bounces at nanog.org<mailto:nanog-bounces at nanog.org> on behalf of owen at delong.com<mailto:owen at delong.com>> wrote:

On Dec 28, 2017, at 18:54, Ricky Beam <jfbeam at gmail.com<mailto:jfbeam at gmail.com>> wrote:

On Thu, 28 Dec 2017 21:05:33 -0500, Owen DeLong <owen at delong.com<mailto:owen at delong.com>> wrote:
If you want to make that argument, that we shouldn’t have SLAAC and we should use /96 prefixes, that wouldn’t double the space, it would multiply it by roughly 4 billion.

I'm saying I should be able to use whatever size LAN I want.

   Sounds like you already are and nobody is telling you that you can’t. It’s a rather silly way to over-complicate your life, but if you want to be on the wrong side of a direcTV commercial, nobody’s trying to stop you.

The routing problem might be real if everyone goes to PI, but I think that’s an unlikely scenario.

Every scenario everyone has come up with is "unlikely". Home networks with multiple LANs??? Never going to happen; people don't know how to set them up, and there's little technical need for it.

   Lots of home networks have multiple LANs today, so you’re patently wrong there already.

Your definition of “amazingly fast is pretty odd... we’ve allocated tiny fractions of 2 /3 prefixes to special uses (multicast, ULA, loopback, unknown, etc.). Beyond that, there’s a /3 delegated to IANA as unicast space for distribution to the RIRs. Of that /3, IANA has delegated a little more than 5 /12s to RIRs. That’s the total of 20 years worth of turkey carving and constitutes well under 1/8th of the address space. Issued. By that measure, we’ve got well over 160 years to worry about runout.

After 20 years of not using IPv6, that's actually A LOT of carving. And if you look at what's been assigned vs. what's being announced vs. what's actually being used, there's a fantastic amount of waste. But nobody cares because there's plenty of space, and "we'll never use it all." (history says otherwise.)

   Given that more than 50% of US mobile traffic is now IPv6, I find it hard to give credence to a claim of “not using”. It’s also north of 40% for US fixed wire line traffic.

   As I said, I don’t doubt that we may eventually run out. However, I doubt anyone alive today will still be alive when we do.

   Owen