The "internet of things" exists already and has done for years, and it works fine over IPv4.

IoT devices talk to a "meet in the middle" server, not directly to the end user, for multiple reasons:

1. To allow communication between devices which are behind a firewall. Even IPv6 home networks generally have an "allow-outbound-only" policy on the firewall
2. Because devices tend to change their addresses (especially on IPv6)
3. So that they can manage your devices 24x7, when the user is not online
4. So that they have full control of the user experience (and collect lots of valuable data that they can sell on)

The fact that it also enables NAT is a bonus, but not the fundamental reason why they use this architecture.

In reply to a post by pluralist:

The "internet of things" and self-driving cars will need IPv6.

Nope, and nope.

In reply to a post by jchamier:

I saw an interesting article in the last week that realised IPv6 “may not be needed” given things like NAT and CGNAT have happened to keep v4 working.

That depends on how "need" is defined. The internet was never going to stop working if IPv6 wasn't universal, but IPv4 address exhaustion undoubtedly causes problems.

In reply to a post by Oliver341:

That depends on how "need" is defined. The internet was never going to stop working if IPv6 wasn't universal, but IPv4 address exhaustion undoubtedly causes problems.

The question, and balance, is if the "problems" of working round exhausted v4 is less work than the confusion v6 causes for many people. I long said that we needed a mid-ground, as the un-memorable IPv6 addresses is a problem themselves.

In reply to a post by jchamier:

I long said that we needed a mid-ground, as the un-memorable IPv6 addresses is a problem themselves.

Only for geeks like us. At least 95% of the world population haven't a clue what IPv4 and IP6 are, and never will. Most people I know don't even know what the acronym URL means or what it does. They just know the published "word(+ number)" address for websites they might visit.

Re the IoT and self-driving cars:

Yes the first has existed for quite a while now. But we can expect a hyperbolic increase in the number of "things" over the next decade. As for self-driving car similar applies, bearing in mind they are fast moving and need far better internet links than we now have.

I expect you are well aware of sat-nav and dash-cam speed indicators lagging several seconds behind actual. GPS is of course not going to be the answer, but whatever the answer is we don't yet know. Several interesting links are given here.

5G or higher becomes essential, and "masts" will need to be every few metres.

Then add in delivery drones above and landing amongst us, and maybe flying taxis.

Is IPv4 plus CGNAT going to handle that lot?

+++++++++

Street power failures will be fun. There's a 39-second video in one of those links. See where the driver's hands are. Think of the drivers' reaction times if the cars in the area suddenly revert to our current self-driving tech.

In reply to a post by Oliver341:

In reply to a post by jchamier:

I saw an interesting article in the last week that realised IPv6 “may not be needed” given things like NAT and CGNAT have happened to keep v4 working.

That depends on how "need" is defined. The internet was never going to stop working if IPv6 wasn't universal, but IPv4 address exhaustion undoubtedly causes problems.

Exactly. How will IPv4 plus CGNAT work once the number of servers + net routers + websites approaches 4,294,967,296.

In reply to a post by pluralist:

Re the IoT and self-driving cars
…

I was behind this Ford Mustang from Wayve waiting at the lights this afternoon. Didn’t clock at first it was on a training mission in East London 😂

Hope this bet pays off for Masayoshi Son!

In reply to a post by pluralist:

Exactly. How will IPv4 plus CGNAT work once the number of servers + net routers + websites approaches 4,294,967,296.

I'd happily share an IPv4 address, my IP address needs are fully satisfied by IPv6 now. I don't think the SR203 is on Sky's roadmap to be enrolled in MAP-T though.

I don’t see the need for self driving cars to have internet access. They may need to have self organising network between them on a stretch of road

Remember the people whom said we needed IPv6 mostly made up the use cases before these other things were designed.

Now you are being fanciful.

No internet connecting them. Bluetooth impossible in most circumstances at junctions. Similarly radar. A completely new method of communication required.

I doubt if AI will give us multi-vehicle telepathy.

Mind you, Musk might give us human telepathy for the masses, not as present for the few.

Quantum entanglement to avoid crash entanglement?

Edited by pluralist (Mon 11-Nov-24 00:05:41)