Last IPv4 Addresses Sold: What Now?

Forgot to say, "For questions you might have, I can try to help, but for specific ISP plans, you should call your ISP."

2 observations.  While most of the Internet is not accessible via IPv6 some of the big boys have already converted so you can reach them either way (Google and Facebook are a couple that come to mind).  It will be a slow process until the break point - and that will come now that the IPv4 pool is gone.  A massive change like this is never easy, but as you note, necessary.

And second, the mistake of IPv4 was not learned.  While the address space has increased, so has the waste.  While the number of IP addresses per person seems astonishing, for now, most will never be used because of that waste.  That means in another 30 years we will be back to square one on IP addresses (as everything from your car to your cat will need one).  But do I care?  I do not plan on being employed in 30 years, so not really.

This has been a long time coming, the ISPs should be prepared by now. Running small-scale tests at this hour is being late to the party.

Same here.

In thirty years? Nah ....... it'll be too scary.

Only part of the story. The ISP's have to have equipment which works, and software that balances loads between servers, and that doesn't work reliably yet. Also, who's doing the certifying of the hardware and that has to be verified. No one really knows how to work with the new protocols and IPv6 addresses and Security software isn't adapted yet for IPv6.

One thing is that there are millions of extra IPv4 addresses that are unused which were given away to IBM, the U.S. Postal Service, the airline operations support company SITA, Prudential Securities, pharmaceutical giants Merck and Eli Lilly, the Massachusetts Institute of Technology, Apple, Xerox, AT&T, Level 3 Communications, General Electric, Ford Motor, and Halliburton.

Maybe 14-28% are in active use. The world used 319,000,000 last year alone. Also, trying to get them back isn't easy as how do you ask? "Can we have the url's you didn't use (but might very well be able to sell) back? As they try to get IPv6 up and running, the IPv4 addresses will become more and more expensive to obtain.

So the transition is going to be very awkward and slow.... and practically the ISP's really had no incentive to invest in unproven equipment and software for something which was then, many years away.

Yea, MIT - did you hear they requested - and got - a /30 for the IPv6????  They must plan on having a LOT of robots!

Why can't they just make a sytem that allows for incremental increase? Don't tell me I to start recycling IP adresses too?

No surprise judging on how they did with the /8.

Because no one can predict the future, mistakes cost money and because an address is either reachable or not.... If you have IPv4 and the site you need is IPv6, you won't get there... it won't matter to you that you can get somewhere else.

Look at it this way: You want to call Mr. Jones, but he has a 13 digit number and you can only dial 10. Is talking to Ms. Green (who has 10 of those 13 numbers in the right order) the same?

And there is no way of making some software that can mediate between the two? Not my area of expertise but last time I wanted to talk to Jones I just called Mr.Cloud and he translated everything for me. But I defer to your wisdom.

They do have work arounds now if you want to run IPv6 internally.  But Like Doc said, if you want to communicate with the rest of the world, you still mostly have to use IPv4.  The problem with using an open ended addressing scheme is computers do not understand variable length addresses.  They pad with 0s.  So IPv6 is now using 128 bits - but again, with all the waste, that will be exhausted in time (but then they can start doing what they did with IPv4 such as NAT and Classless Addressing).

There is a method to their madness.  Right now, equipment is not set up to handle the number of addresses that are possible with IPv6 being used the way IPv4 is being used.  But every year, routers get more memory, and faster processors, so when they do run out of addresses, the equipment will be able to handle more routes (without causing a bottleneck for the lookup time).

The only thing sure in life and technology is change.  When IPv4 was created, the state of the technology then would not allow CIDR routing either.

^ correct. Perhaps seanw3 is referring to "tunneling" software?

http://www.tcpipguide.com/free/t_IPv6AddressSizeandAddressSpace-2.htm

Just so you get an understanding of the vastness of the IPv6 address range

Tired canards aside, IPv6 is ridiculously overengineered.  There is enough space in IPv6 to allocate 20 IP addresses to every molecule in every human being's body.

In exchange for this, we're losing the ability for humans to use IP addresses at all.  People can remember 192.168.21.17.  People won't remember 4d8c:77ee:0034:128f:f000:3812:a9d3:34ee.

There's a reason why we don't let the engineers design the UI...

Now the world ends.

^ no reason to remember that address, or any other. Jafo pointed up the absurdity of those who said "We'll never need more than...".

We always end up needing more.

No reason to remember that address?  How do you troubleshoot when DNS is down?

Anyway, what I'm pointing out is that "we'll always need more" is a fallacy.  There is such a thing as enough. 

You write it down and store it in a phone or other device. The thing is, I hate cluttering my desktop and my brain. Unfortunately "enough" led to IPv4.... here we go again....

IP addresses should just not be used; just get the IP from the DNS server. Trouble shooting should be done by your operating system anyways.

------------

I really wich MS would restructure its networking API so that it would be agnostic about the IP version. Then future switches would be easier as older apps would stiil work (taking one part of the steps of change overs out of the picture, as the OS would just need to be updated).

More will not be needed, unless we make nanites that need an IP address each or move to other planets and have FTL long distance communications between them, but those are possibilities that should not be ruled out.

And it is a power of 2, which for computer stuff is more effecient. That is why is so big. They could have done a 64 bit value but changing over is such a pain...

"Anyway, what I'm pointing out is that "we'll always need more" is a fallacy.  There is such a thing as enough. "

Let's just concede that it's NOT a 'fallacy' but an OPINION, and a flawed one.

Unless we're into ZPG,

....and stagnation.

Go ahead. Tell us exactly how many IP addresses we'll EVER need. Then we can just go to the point and say 'we have enough'

Can't do it? Then better to have overhead room than to come up short.

It's just like the Y2K thing.

^Correct. Also exactly what I wrote in the OP.

Any technology that looks remotely like what we have today will never fill IPv6.  Suppose each human being needs 1000 devices, each which need 10 addresses, the world population increases to 100 billion, and 99/100 addresses that are alocated go unused due to waste of one form or another.  That's still only 10^17, or just over 2^56 addresses.  This is still considerably less than 1% utilization of a 64-bit range.  Or, to put the absurdity another way, let's say we want IPv6 to last for the next one million years.  We could safely issue one trillion IP addreses (that's 10^12) per second in the intervening period without depleting it.  Alternately, a 64-bit option would only be able to handle an average of just under 600000 IP addresses issued per second over the next million years. 

Until we start handing out IP addresses to nano-bots, the idea of running into the limitations of even a 64-bit range is ludicrous.  Perhaps one day in the distant future this may be the case, but it will be devices that look and act nothing like the devices of today that run into these sort of limitations.  For this reason, I find it significantly more plausible that IPv6 will be obsoleted for unrelated technical reasons long before it even comes close to filling a 64-bit range.  Feel free to come back and tell me "I told you so" if we ever do get that far, but I sincerely doubt you will get the chance.

That said, I have nothing against over-engineering in and of itself; it's not like it's difficult to do, and the comedy of the sheer scale of the address space is well worth it as far as I'm concerned

On the other side of the spectrum, I find it disgraceful how everyone has waited until IPv4 is actually depleted to make the transition.  This has been coming up for years, it's not like anyone in the field can claim surprise.

...Never need more than 640k.

Before that it was... Impossible to travel faster than sound.

Now it's... Won't ever use all the IPv6 address space in a million years.

What will be next?

Oh, right... Can't travel (as in physical transportation) faster than light.

Which will be proven false, first?

Even though I probably won't be alive to see it, I'd bet on IPv6.