Today's ARIN estimated depletion date:

DNS when turning on IPv6


I did some interesting IPv6 and DNS testing the other day with a customer. I decided to put a little blog post together on my findings. You can read the post at the Secure64 blog about DNS



Tags: , ,

13 Responses to “DNS when turning on IPv6”

  1. Mindbuilder says on :

    Sorry if this is off topic, but the traffic is so low now I thought it wouldn’t matter. It seems Huston is predicting RIPE depletion in just nine days from now on September 21, but the prediction here is not for about five more months. Is Huston way off? Why the difference?

  2. Andrew Bower says on :

    Mindbuilder, if you look at RIPE’s chart it suggests Huston is on the money!

  3. Andrew Bower says on :

    Link got munged there:

  4. rb12345 says on :

    For what it’s worth, the Huston prediction may still be slightly optimistic considering the apparent last-minute rush for addresses. I’d guess that RIPE will run out early next week.

  5. Andrew Bower says on :

    And today they’ve gone!

  6. ipv4depletion says on :

    Thanks, Looks like I have to adjust my predictions here a little….

    I must have some reserve space that I’m counting that should be there.


  7. ipv4depletion says on :

    The strange thing is that some space that are allocated does not show up in the RIPE file.

    A few examples are – that does not show up in the latest delegated-ripe file.
    however, if you do a whois it says that this network is delegated to EWE tel in Germany.

    Same thing with Not mentioned in the statistics file, but whois says Jump Internet in Romania.

    There are plenty of those examples. Somebody from Ripe that can explain?


  8. rb12345 says on :

    I’m not from RIPE, but whois for gives a netname of “DE-EWETEL-20120914″, so I’m guessing it’s newly-allocated address space and will show up in the next daily update of delegated-ripe. is similarly named.

  9. ipv4depletion says on :

    Those are only two examples. there are plenty of other (enough for RIPE to last until next year).

    For example, what is – and why can’t they use that?, ok when did they do that? why?


  10. Grahm says on : is reserved for was moved to Afrinic in 2005. Why? Because RIPE african members became Afrinic members together with their space.

    For years you were saying that you knew about some secret ranges that RIPE lost but should use and were refusing to name them. And now, when it’s too late, you started revealing your findings. Nice.

    BTW, where were you getting your data from?

  11. rb12345 says on :

    Grahm: I expect the real question about the temporary space /13 was this: why retain the /13 temporary address pool when RIPE has nothing left to allocate/assign except that and the final /8? The temporary pool is only allocated for short periods (1 month or so for /14, longer with renewals for /13 or smaller), so these can be reclaimed quickly when RIPE eventually needs the addresses. Using them now would only gain you an extra day or so in the final rush we saw and lose the ability to assign address space to conferences and the like post-depletion.

    I could well be wrong, but I expect the data comes from the RIPE delegated prefixes file (mirrored at, but there will be an official link somewhere at RIPE) and the IANA equivalent. If you know the RIR got a /8 from IANA, you can then look for gaps in their allocation records. Of course, that falls down a bit for the reassigned-to-Afrinic addresses and the temporary pool, and there seem to be gaps in that file regardless ( for Sky Broadband).

  12. Grahm says on :

    rb12345: “delegated” does not have free space. For 1.5 years already there is also an “extended” RIPE file, with free space listed.

    It’s a good thing to find missing bits and tell RIPE that they lost them. But maybe better do this before they switched to /22 per cust.

    Why is this a gap?

  13. rb12345 says on :

    It wasn’t meant to be a gap, but I think I simply missed that line in my rather quick search through the file. (Thanks for finding the links to the original RIPE source too!)

    My personal assumption though is that there are no missing bits after all, just reassigned space like the Afrinic example.

Leave a Reply