U.S. Will Run Out of Internet Addresses This Summer

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
If only we'd had known about this issue sooner, maybe we could have saved the internet. :rolleyes:

The shortage isn’t as dire as it may appear. The 4.3 billion limit applies to IPv4. But IPv6, approved in 1998—IPv5 never caught on—allows for a mind-boggling increase in addresses to 340 undecillion, or 340 followed by 36 zeroes, enough to assign an IP address to every atom on Earth.
 
No such thing. People have been saying this forever, but thanks to NAT it really doesn't matter.
Never heard of IPv5 though.
 
I wonder how long it'll be until everyone in the US has a permanent IP address.
 
Interesting. I never realized there was an IPV5 that never caught on.

I always just assumed that they used an even-odd development/stable version numbering scheme, similar to how the Linux kernel used to operate.
 
Meh, It' won't be the end of the world.

Short term this will just create a market for buyng and selling of IPV4 addresses between those who ahve them issued. Those who hae greater need will be willing to pay more for them.

long term this increase in cost will shift some to more heavily use of NAT and others to adopt IPV6.

Eventually we'll all be on IPV6.

IMHO, IPV6 is the way to go, but it is too bad the human readability is more convoluted than with IPV4.

I would have preferred to add one or two more blocks of 256 numbers to IPv4 and call it a day.
 
If Direct Access is any indication of how well IPv6 is going to work, then screw IPv6.
 
Zarathustra[H];1041601292 said:
Meh, It' won't be the end of the world.

Short term this will just create a market for buyng and selling of IPV4 addresses between those who ahve them issued. Those who hae greater need will be willing to pay more for them.

long term this increase in cost will shift some to more heavily use of NAT and others to adopt IPV6.

Eventually we'll all be on IPV6.

IMHO, IPV6 is the way to go, but it is too bad the human readability is more convoluted than with IPV4.

I would have preferred to add one or two more blocks of 256 numbers to IPv4 and call it a day.
Or maybe screw hexadecimal and just represent the 2-byte blocks in 10-base like IPv4?
 
Zarathustra[H];1041601390 said:
Interesting.

I wonder who the users are. I've certainly never encountered it in the wild...

if I go to whatsmyip I see an ipv6 address. This is with Comcast Business in Illinois. I do have a block of static IPs, but this is what i see if I connect through their gateway without using one of those.
 
Or maybe screw hexadecimal and just represent the 2-byte blocks in 10-base like IPv4?

You would wind up with a pretty long dotted decimal representation. A typical IPV6 address would look something like this:

128.91.45.157.220.40.0.0.0.0.252.87.212.200.31.255

I don't understand why they felt the need to go all the way to 128bit addresses.

Sure, it's good to preempt increases, so you don't have to keep changing a system, but it makes them impractically long and difficult to memorize.

I would have been happier with an incremental approach. Maybe 48bit at first, followed by 64 if necessary.

~3.4×10^38 addresses is overkill to the point of lunacy at this point. I mean it's more than 7.9×10^28 times more than IPV4...
 
if I go to whatsmyip I see an ipv6 address. This is with Comcast Business in Illinois. I do have a block of static IPs, but this is what i see if I connect through their gateway without using one of those.

Neat. Mine both at work and at home is still IPV4. I haven't even bothered enabling IPV6 on my router or local network as I find it to just be a nuisance.

Took me long enough to wrap my head around subnets in IPV4 CIDR notation, I don't want to have to learn everything over again :p
 
Once my ISP finally does go IPV6 (it's bound to happen at some point) will I be able to keep my local network all on IPV4 and just have the NAT handle transation to an IPV6 WAN, or am I going to be forced to transition my local machines to IPV6 in order to have them communicate fully with external IPV6 servers?
 
No, IPv6 addresses look like MAC addresses. Delinated by colons.

Like this: 00:01:60:8000:01:88
 
The original forecasting could've been better but I understand their struggle with limited compute resources back then. If they had gone with 16-bit instead of 8-bit per octet such as FFFF.FFFF.FFFF.FFFF we would have 18,445,618,199,572,250,625 instead of 4294967296 addresses which would make the shortage a non-issue now but also into the distant future while retaining the easy familiar format.
 
Just for simplicity sake I'm hoping that when IPV6 is forced on me externally, I'll be able to set up something like this without any loss of functionality:


Local network IPv4 only. IPV6 completely disabled on all hosts/clients -> NAT -> WAN IPV6.

I'm suspecting this won't be possible though, as I can't imagine a NAT can completely translate the full range of IPV6 addresses into IPV4 format.
 
The original forecasting could've been better but I understand their struggle with limited compute resources back then.

Yeah, but a 128bit address space was overkill any which way you look at it.

to put things in perspective:

One main distinguishing characteristic of IPv6 is that that there are 2^128 addresses. That’s 3.4028x10^38. To put that in perspective, there are only about 10^19 grains of sand on all the beaches of the world. If you assigned an IPv6 address to each grain of sand, each one could have more than a billon billion addresses.
 
"enough to assign an IP address to every atom on Earth. "

Well, the way the surveillance state is increasing this might not be far off. :D
 
They've been saying this for years now. Its ridiculous clickbait for the uninformed.

NAT fixed a lot of issues and freed up a bunch of addresses. You can be behind a router with one public address and thousands of private addresses behind it.

Then when enough of the IPv4 addresses were being eaten up despite NAT, then IPv6 started to take off and relieved another huge chunk of addresses.

Even more people connected to net being assigned IPv4 or IPv6 addresses depending on the ISP, what was left ate up some remaining IPv4 addresses.

More people will migrate to IPv6 per their ISP, and the whole wheel keeps turning. No doom and gloom to be found.
 
Or maybe screw hexadecimal and just represent the 2-byte blocks in 10-base like IPv4?

Why is 65342.31302.5230.99 is easier than FF3E::7A46::146E::0063 (or 63). I could be wrong, but I don't think most understand (or care) what 192.168.1.1 means.

My guess is that companies care because of legacy equipment, which I assumed had been upgraded, since we've been talking about IP addresses for years.
 
Zarathustra[H];1041601417 said:
You would wind up with a pretty long dotted decimal representation. A typical IPV6 address would look something like this:

128.91.45.157.220.40.0.0.0.0.252.87.212.200.31.255

I don't understand why they felt the need to go all the way to 128bit addresses.

Sure, it's good to preempt increases, so you don't have to keep changing a system, but it makes them impractically long and difficult to memorize.

I would have been happier with an incremental approach. Maybe 48bit at first, followed by 64 if necessary.

~3.4×10^38 addresses is overkill to the point of lunacy at this point. I mean it's more than 7.9×10^28 times more than IPV4...
For the same reason people back in the 8-bit days thought 64-bit CPUs were beyond necessary.
 
Zarathustra[H];1041601450 said:
Just for simplicity sake I'm hoping that when IPV6 is forced on me externally, I'll be able to set up something like this without any loss of functionality:


Local network IPv4 only. IPV6 completely disabled on all hosts/clients -> NAT -> WAN IPV6.

I'm suspecting this won't be possible though, as I can't imagine a NAT can completely translate the full range of IPV6 addresses into IPV4 format.

in theory you dont need nat with IPv6 every rock can have an IP
 
there is also a lot of waste in the IPv4 space

take HP they have 2 or 3 FULL class As they are sitting one
 
in theory you dont need nat with IPv6 every rock can have an IP

True, but if my local network is IPV4 then I would need some sort of translation.

It looks to me you should be able to create some sort of NAT translation that enables IPV6 machines to make connections to IPV4 machines, but not the other way around, as IPV4 clients would have no way of addressing the overwhelmingly large amount of IPV6 IP addresses.

Also, even if you don't NEED NAT, I would kind of want to do something to close off my local network from the outside world and assign my own addresses to things. Just firewalling a bunch of public IP:s just doesn't feel right to me. I want to be able to customize my boxes which I have memorized based on the last digits of their IP and knowing what private IP range I have chosen.

In theory - however - you won't need a router anymore. Just a switch. (And unless you are feeling adventurous, some sort of firewall)
 
For the same reason people back in the 8-bit days thought 64-bit CPUs were beyond necessary.

IPV6 has an address space able to provide more IP addresses than there are atoms on planet earth. In any conceioveable future usage scenario, even if we give absolutely EVERYTHING it's own IP address, that is still a few billion billion times more IP addresses than could ever under any circumstance become useful, unless you are going to make single atom computers out of every atom on earth... :rolleyes:
 
You don't have to input all the numbers in ipv6, if your setting up a network now days you are probably just going to adding about 1 or 2 extra blocks, which would be no different than the proposal to extend ipv4 by 2 blocks. Also with more addresses possible then various blocks at the higher level will probably start to be easier to remember because they will be more consistent. Sort of like how its easy to remember the area code in a phone number even though technically it adds more numbers, you just know everyone has the same area code.

Another thing I would like to see with ipv6 having a lot more space is companies that issue blocks giving people the option to choose their own. So maybe you get to choose the last 2 blocks and you can make into numbers or letters you find easy to memorize.

I would also like to see ipv6 allow more area based IP addresses so like everyone in the USA has the same first block just like everyone has the same country code in a phone number. Then each state gets the next block, after that its broken down by company.

TLDR don't worry about IPv6 it brings a lot of options, if people think that the long numbers will be a significant barrier to moving forward the best thing for all of us is to see ipv4 run out of space so they can be forced to work on solutions to all these problems.

And I definitely disagree with the notion of creating baby steps we have to keep coming back an revisiting every so many years. Move to IPv6 lets get it over with, the longer people use something the more they will understand it and confusion or issues with memorization will become less of a problem. The longer we delay the more garbage gets released that holds us back even longer.
 
Stopped reading at "IPv5 never caught on". Its clear the author doesn't know shit and I'll be made dumber by actually reading this "article".
 
Certainly hope IPv6 doesn't catch on. Tossing out the TCP/IP stack and having every device on a public connection raises a whole bunch of security concerns.

I'm friends with a senior IEEE member and pretty much was in the room when IPv4 was born. And he does not like it one bit.
 
IPv5 was never an addressing scheme like IPv4 or IPv6. It referred to an experimental streaming protocol: http://programmers.stackexchange.com/questions/185380/ipv4-to-ipv6-where-is-ipv5

I also love how in every topic like this there'll always be a few people saying that NAT is a good thing, and that we can squeeze a lot more blocks of unused addresses out of large corporations and governments. The reality of the matter is that we will run out. Soon. Not this year, not next year, but by the next decade for sure. Not instantly either, but it'll become increasingly more difficult and more expensive to get a new address.

Also, NAT is completely evil. It screws up routing and goes in against the basic principle of the internet, in that you can route anything to anywhere as long as you know the start and end IP address. While IPv6 isn't perfect, it is a major improvement on IPv4, and we should have embraced it more than a decade ago already.
 
You do realize that H stands for hemorrhoids?

Originally that was the case, but now they sell products not intended for hemmorhoids under that same brand name too, like - for instance - hydrocortisone butt cream.
 
Zarathustra[H];1041601949 said:
True, but if my local network is IPV4 then I would need some sort of translation.

It looks to me you should be able to create some sort of NAT translation that enables IPV6 machines to make connections to IPV4 machines, but not the other way around, as IPV4 clients would have no way of addressing the overwhelmingly large amount of IPV6 IP addresses.


It's way more than just addressing. The IPv6 protocol does use the TCP/IP stack protocol and currently there is no way to translate IPv6 to IPv4 (or vice versa). Which is one of the main reasons why it's taking IPv6 so long to become the new standard.

Having two infrastructures on different protocols is way too expensive and there aren't enough people to switch over to justify the cost. It literally would have to come down some sort of federal regulation that IPv4 HAS to switch to IPv6 by (Let's just say) Jan 1, 2026.

There there will be A LOT of push back on that.
 
It's way more than just addressing. The IPv6 protocol does use the TCP/IP stack protocol and currently there is no way to translate IPv6 to IPv4 (or vice versa). Which is one of the main reasons why it's taking IPv6 so long to become the new standard.

Having two infrastructures on different protocols is way too expensive and there aren't enough people to switch over to justify the cost. It literally would have to come down some sort of federal regulation that IPv4 HAS to switch to IPv6 by (Let's just say) Jan 1, 2026.

There there will be A LOT of push back on that.

I was under the impression that NAT64 existed to make the IP stack between the two, and that because of this addressing was the only remaining problem, and that IPV6 machines could address IPV4 easily due to IPV4 being a subset of IPV6 addresses, but the opposite wouldn't be so straight forward.
 
Back
Top