The problem isn't the number of addresses, it's the address space fragmentation. The IPv4 global routing table is already 3x the size it should be. At the point where you're just giving every device an arbitrary number, the addresses cease to function as addresses.

LISP is meant to solve this problem, but it's yet to be deployed outside of the test network. It could also be used to address exhaustion to some extent by allowing the use of prefixes longer than /24 on the internet, although that is not the intention.

https://www.lisp4.net/

https://neverthenetwork.com/notes/lisp/

Full disclosure, the second link is my blog.

I guess this would give us a few years of breathing room by letting us fragment ipv4? Does anyone support this other than Cisco (ie Broadcom/Cavium/Intel/Barefoot)?

Really it's meant to solve the fragmentation that already exists. Cisco developed the standard, but it's completely open.

I believe Cisco is the only large vendor that's implemented it, but there are Linux and BSD implementations, which should make it easy for any Linux or BSD based NOS to implement it in the future.

Honestly this problem will probably be mostly solved in hardware with larger FIBs/TCAMs

Edit - I should clarify the problem I'm referring to here is fragmentation, not exhaustion.

Especially if you stop allocating TCAM memory for the much larger v6 addresses. I hate to think how much memory is sitting out there in routers, configured for 128 bit v6 addresses and never used...

You only need 64 bits for routing v6, and since the address space isn't fragmented the routing table ends up at a similar efficiency level to the v4 one - probably better once you take into account not needing CIDR.

Indeed, until now there has been no financial incentive for those companies to sell the IPs. At a certain point demand and willingness to pay rise making sellers interested in selling. Or maybe IPv6 will happen instead. The problem is IPv6 requires a lot of very complex inventive alignment, whereas an IPv4 market will just result in a well understood price/scarcity relationship.

The going rate for an IPv4 address is at least $10 each. Selling a /8 block would be a nice $160 million profit.

Ouch! That makes Stanford's voluntary surrender of their /8 block back in 2004 extraordinarily expensive! Think about how many scholarships that could have paid for!

That price is based on what Microsoft paid for Nortel's block in 2011.[0] I think there was another purchase of IP addresses since (by Amazon, maybe), but I'm not able to locate a price.

[0] https://www.networkworld.com/article/2228854/microsoft-pays-...

DuPont sold off their 52.a.b.c to Microsoft and Amazon 5 years ago.

I've been sitting on a /23 for almost 10 years for this day. Who wants to lease my block

First reaction (after looking up that this is 8 million addresses): wow, nice!

Second reaction (after the "$10/IP" part clicked): oooh, this person is going to have a lot of money soon o.o

But you don't have any contact info in your bio!

(I'm in no position to seriously respond - I'd only want one or two addresses to play with)

Er no, a /23 means they control the last 32-23 = 9 bits, so that's 512 addresses. So /if/ they sold that for $10 per address which is plausible that would be about $5000.

._.

Thanks.

It's well over 10 today.

AT&T is an ISP, with millions of customers. I'd expect them to need a /8, or more.

AT&T (the ISP's) allocations are to AT&T Corp, not AT&T Services.

AT&T's 12/8 is used for end-user address assignments by AT&T Services. See https://bgp.he.net/AS7018#_prefixes for their announcements.

Most of their broadband subs are from legacy SBC (ameritech, swbell, nvbell, pacbell, snet). They got all their allocations separately. None of them came out of 12/8 or 32/8 traditionally.

Apple and USPS are not ISPs. It doesn't seem like they need a /8 though.

Comically when Apple would setup a store they'd ask their isps to provide an allocation rather than dip into their allocation.

From a global routing perspective this makes sense. Otherwise every store has to advertise its netblock on BGP. Not impossible, but it bloats the routing table unnecessarily. Better for Apple to keep those addresses on its campus and let the stores connect like regular Internet endpoints.

Even better would be for Apple to return most of those addresses and only use a handful of /16s for its hosted services like any other business and switch the internal nets to RFC1918.

So, in the speeds ipv4 was handed out before they ran out, how long would two /8s last? Weeks? Months? You'd be able to slack on moving to ipv6 for those weeks, months. Whee.

The last two /8s were handed out in January of 2011 so quite a bit longer than weeks or months.

https://en.wikipedia.org/wiki/IPv4_address_exhaustion#Addres...

Entire networks of computers can share a single public IP address so the 33.5 million public IP addresses supplied by two /8s could last quite a long time. Or we can let Ford sit on them.

So given the rate at the time,

https://en.wikipedia.org/wiki/IPv4_address_exhaustion#/media...

with some 1M ips given out PER DAY, how long would two /8s last? 4 weeks, or 33 days. That is not "quite a bit longer", it is literally what I suggested.

It is only now, long after the crunch that thinking entire networks would share a single ip (hopefully never running any SIP,bittorrent or something that needs more than one port simultaneously) or the entire network would quickly run out of the 64k ports usable...

So, if Ford and whoever had that second net did give it back when it started to get scarce, all of 33 days is what we would have "won". Good margin for writing up that ipv6 migration plan I guess.

Ford? is there a listing of what companies own the largest blocks?

EDIT: just found it https://en.wikipedia.org/wiki/List_of_assigned_/8_IPv4_addre...

I don't think Apple will ever give up their IPv4 block unless forced by legal means of some sort. They use the massive IPv4 block internally if I recall correctly.

I could see most others doing it though.

Yes I heard the same thing. Must be really convenient having a publicly addressable (probably not publicly routable) IP address for every machine internally.

Yes, it does make a lot of stuff much easier. Which is why it is so painful to see IPv6 mired in the mud for year after year.