I noticed discussion in the "IPv4 Depletion Party" thread regarding RIPE's depletion. I want to catalog everyone's predictions here.
Early in that thread I noticed that the estimate that Geoff Huston had on his website was unrealistic. And indeed we are two weeks past that estimate, and we still haven't run out. He did update his estimate before we reached the estimated date.
In the same post I did an extrapolation based on 11 weeks on consumption, as published on the RIPE website. And that gave me October 10th. Right now the same method gives me an estimate saying October 14th. But that is a Sunday.
My guess right now is that once the folks at RIPE arrive at work on Monday morning October 15th, they'll find they have received more requests over the weekend than can be fulfilled. And we'll see an official announcement on that Monday that RIPE has reached the last /8 and has started rationing.
Wow, I suck at prediction. I took the past twelve weeks and did two regressions on it:
Linear: February 14, 2013
Quadratic: March 9, 2013
Quote from: wswartzendruber on August 14, 2012, 11:10:42 PM
Wow, I suck at prediction. I took the past twelve weeks and did two regressions on it:
Linear: February 14, 2013
Quadratic: March 9, 2013
That's about the date you'll get if you assume all the remaining addresses are handed out according to the current policy. But they won't. The last /8 is handed out according to a different policy, which is the rationing I was referring to.
As long as there is more than a /8 left, LIRs can get IPv4 addresses to keep their IPv4 operation going. Once there is only a /8 left, LIRs can only get tiny allocations of IPv4 addresses intended for those cases where you need a small number of IPv4 addresses for the transitioning from IPv4 to IPv6.
The rationing should ensure that there will always be a small number of IPv4 addresses left for such transitioning mechanisms. If the last /8 runs out before IPv4 usage start declining, we'll be in even more trouble.
What we are trying to predict is when the rationing happens, thus when there is only one /8 left. Now if you redo your calculations with that in mind, I'm sure you'll end up in mid October like the rest of us.
I can't believe I missed that. So I forgot to subtract 2^24.
I went back and did it again:
Linear regression: December 31st
Quadratic regression: January 9th
This doesn't factor in any change in rate of consumption.
Quote from: wswartzendruber on August 15, 2012, 07:22:20 PMI went back and did it again:
Linear regression: December 31st
I just now tried doing a linear regression using this tool (http://www.easycalculation.com/statistics/regression.php), and I got 15th of October.
I wonder how that one guy came up with last month for a time.
Hi there
RIPE NCC has Approximately Four Million IPv4 Addresses Before Reaching Last /8;
http://www.ripe.net/ripe/mail/archives/ripe-list/2012-September/000615.html
This put them into phase 1;
http://www.ripe.net/internet-coordination/ipv4-exhaustion/last-8-phases
The graph is going to be updated daily;
http://www.ripe.net/internet-coordination/ipv4-exhaustion/ipv4-available-pool-graph
Geoff Huston has the current exhaustion date at October the 4th;
http://www.potaroo.net/tools/ipv4/index.html
Regards,
Rob
Hi there
According to some, RIPE's IPv4 depletion is going to occur during RIPE65;
https://ripe65.ripe.net/
There will be webcasts.
Regards,
Rob
Allocations have been speeding up a bit. During the last week there was a decrease from 0.31 to 0.15 blocks. They could run out next week.
Hi there
More graphs;
http://www.ripe.net/lir-services/resource-management/tools-for-lirs/reponse-time-ipv4
At the current rate RIPE will be depleted at the 19th or so.
However, we want to announce our party at least a few days ahead, so it will probably be at the 28th.
Regards,
Rob
Quote from: sput on September 11, 2012, 07:09:22 AMAt the current rate RIPE will be depleted at the 19th or so.
I'll be celebrating that date regardless. I did notice half a year ago, that it was a possible date for RIPE to run out.
Looks like last Friday was the winner.