Posters of the IPv4 and IPv6 internet as of Jan 1, 2023 • Hacker News Board

Re: Posters of the IPv4 and IPv6 internet as of Jan 1, 2023

#101

Post by IcePic » Tue, Jan 03, 2023, 9:23 AM UTC

dtaht wrote: post #99

themoonisachees wrote: post #68

Earlier quoted context omitted.

240/4 ("class E") addresses were intended a buffer for future allocation, but they can't use it for global unicast because a lot of equipment considers any IP above 239.255.255.255 to be invalid. There have been RFCs to make it formally global-unicast and local-unicast but none have succeeded because of the above-mentionned poorly programmed equipment. That said, if you know your equipment can handle it like some net…

There have been multiple attempts at making 240/4 globally useful. It pretty much just works on most gear today (except windows), but the resistance by the IPv6 crowd to opening up 1/16th of the ipv4 internet for general use, astounding. There have been multiple internet drafts that did not make it through the relevant committees. Example: https://www.ietf.org/id/draft-schoen-intarea-unicast-240-03....

This makes it sound as if the ipv6 crowd would be larger than "the windows crowd". Perhaps start with convincing every windows user first, then "the ipv6 crowd" should be easily convinced.

Re: Posters of the IPv4 and IPv6 internet as of Jan 1, 2023

#102

Post by N19PEDL2 » Tue, Jan 03, 2023, 12:02 PM UTC

I noticed that 240.0.0.0/4 is reserved for "Future Use". As IPv4 is going to be replaced by v6 (slowly, ok, but still...), what is that future use supposed to be?

Edit: I just saw that my question has already been answered in this thread.

Re: Posters of the IPv4 and IPv6 internet as of Jan 1, 2023

#103

Post by sgjohnson » Fri, Jan 06, 2023, 1:45 PM UTC

preisschild wrote: post #76

walrus01 wrote: post #10

Earlier quoted context omitted.

As an example of how big ipv6 is, we got a /29 from ripe at a not very big at all ISP, which even if we had handed out a /48 to each and every DIA customer physical location/cpe, would have lasted us at our projected growth rate for the next several hundred years. And that was in addition to the ARIN /32 we already had.

I've got my own homelab /44 prefix for only 15€/yr. /48 would have been fine too, but /44 was the smallest they had.

From some LIRs you can get an ASN + a /40 for life for $50. And even less if they decide to do a sale.

And the /40 would be an actual "ASSIGNED-BY-LIR" sub-allocation that you could also sub-allocate as you see fit.

Re: Posters of the IPv4 and IPv6 internet as of Jan 1, 2023

#104

Post by dtaht » Thu, Jan 12, 2023, 2:57 AM UTC

nerdbert wrote: post #100

adrium wrote: post #91

Earlier quoted context omitted.

It is irresponsible to use (let alone to hardcode) those addresses in the first place. But in any case, fortunately, software can be patched and it has been done so already in 'tricky' cases: Y2K and X.509 UTCTIME are examples. My company operates big networks and it caught us off guard when 44/8 got used on the public Internet. Internal tooling used the space because it was assumed to be non-routable. Assumptions li…

I doubt anything containing 0.0.0.1 is advertised in China (it's definitely not outside of China). Looks like an address someone would use to indicate an error condition. ... which is just another example of the zillions of cases that would have to be dealt with. Sure, each one is probably quite easy. But the sheer number of them is huge, and many of them will only be uncovered after they fail, setting off a frantic…

My principal reason for proposing that 127 be reduced to a /16 instead of a /8 was to help unwind the kubernetes hairball.

I have no idea to what extent 0/8 is being used today. I do see quite a bit of 240/4 in private networks. It's a shame that last may never be publically allocated.