3 ms·

How would making it an extension help? You still have the fundamental compatibility problem that you can't address the wider address space with a narrower addre

How would making it an extension help? You still have the fundamental compatibility problem that you can't address the wider address space with a narrower address.

jeroenhd 5y ago
You'd probably be encapsulating "IPv5" traffic in IPv4 traffic towards known translation routers at ISP networks. CG-NAT with some extra advertisements and routing knowledge, basically. You'd be stuck with an abysmal MTU for "IPv5" though, because you need that wrapper and the world doesn't seem to want an MTU larger than 1500 outside internal networks.
- cesarb 5y ago
  > You'd probably be encapsulating "IPv5" traffic in IPv4 traffic towards known translation routers at ISP networks. That's what 6to4 does. When enabled, it encapsulates IPv6 traffic in IPv4 packets and sends them towards a well-known anycast IPv4 address, and the return IPv6 traffic is sent towards a well-known anycast network, where a router encapsulates it in IPv4 packets and sends them towards the IPv4 address contained within the IPv6 address. I find it amusing that all these "IPv5" proposals either cannot work, or do the same as something which already exists for IPv6 (and in this case, it works; I've used 6to4 for a while).
  - ay 5y ago
    You ought to use the past tense when speaking about 6to4.. As much as technically I think it was a brilliant idea (a whole /48 assigned to each existing IPv4 address owner), it failed because of the tragedy of commons.
    - jeroenhd 5y ago
      One of the problems with 6to4 that make it hard to deploy in well-configured networks is the DNS problem. Google runs a 6to4 DNS but warns that DNSSEC will be broken by design, and clients would need their own translation software on the software side to solve the problem. 6to4 gateways are still a thing, you can traceroute 192.88.99.1 to find the nearest gateway if you have IPv4 connectivity. I've got one seven hops from my device, but I've also got a static IPv4 address so I have little need for a gateway.
  - jeroenhd 5y ago
    You're right about the concept, of course, but 6to4 gateways don't work the other way around. You can't reach an IPv6 network through them, you can only reach IPv4 from IPv6. My silly "IPv5" solution would also have a way for IPv4 sources to send IPv5 traffic upstream. As complex as some may find IPv6, I find it a lot simpler and more independent of routing services than many proposals for alternatives. The weirdest one I've seen involves concatenation two IP headers, giving everyone one network IP for ISPs and such and one client IP for the individual hosts, but that's just replicating IPv6 with extra steps and an even smaller MTU.
cesarb 5y ago
> You still have the fundamental compatibility problem that you can't address the wider address space with a narrower address. There are ways to work around that. All of them (and some more) were tried and/or are in use with IPv6: https://en.wikipedia.org/wiki/IPv6_transition_mechanisms https://en.wikipedia.org/wiki/IPv6_transition_mechanisms
- rcxdude 5y ago
  Those are all ways of allowing ipv6 to address ipv4, (or tunning ipv6 over ipv4 infrastructure) which is fairly easy to manage (getting the reverse connection still involves NAT though). None of them are ways of allowing ipv4 to address ipv6, because that fundamentally doesn't work by the pigeonhole principle.

How would making it an extension help? You still have the fundamental compatibility problem that you can't address the wider address space with a narrower address.

jeroenhd 5y ago

You'd probably be encapsulating "IPv5" traffic in IPv4 traffic towards known translation routers at ISP networks. CG-NAT with some extra advertisements and routing knowledge, basically. You'd be stuck with an abysmal MTU for "IPv5" though, because you need that wrapper and the world doesn't seem to want an MTU larger than 1500 outside internal networks.

cesarb 5y ago
> You'd probably be encapsulating "IPv5" traffic in IPv4 traffic towards known translation routers at ISP networks. That's what 6to4 does. When enabled, it encapsulates IPv6 traffic in IPv4 packets and sends them towards a well-known anycast IPv4 address, and the return IPv6 traffic is sent towards a well-known anycast network, where a router encapsulates it in IPv4 packets and sends them towards the IPv4 address contained within the IPv6 address. I find it amusing that all these "IPv5" proposals either cannot work, or do the same as something which already exists for IPv6 (and in this case, it works; I've used 6to4 for a while).
- ay 5y ago
  You ought to use the past tense when speaking about 6to4.. As much as technically I think it was a brilliant idea (a whole /48 assigned to each existing IPv4 address owner), it failed because of the tragedy of commons.
  - jeroenhd 5y ago
    One of the problems with 6to4 that make it hard to deploy in well-configured networks is the DNS problem. Google runs a 6to4 DNS but warns that DNSSEC will be broken by design, and clients would need their own translation software on the software side to solve the problem. 6to4 gateways are still a thing, you can traceroute 192.88.99.1 to find the nearest gateway if you have IPv4 connectivity. I've got one seven hops from my device, but I've also got a static IPv4 address so I have little need for a gateway.
- jeroenhd 5y ago
  You're right about the concept, of course, but 6to4 gateways don't work the other way around. You can't reach an IPv6 network through them, you can only reach IPv4 from IPv6. My silly "IPv5" solution would also have a way for IPv4 sources to send IPv5 traffic upstream. As complex as some may find IPv6, I find it a lot simpler and more independent of routing services than many proposals for alternatives. The weirdest one I've seen involves concatenation two IP headers, giving everyone one network IP for ISPs and such and one client IP for the individual hosts, but that's just replicating IPv6 with extra steps and an even smaller MTU.

cesarb 5y ago

> You still have the fundamental compatibility problem that you can't address the wider address space with a narrower address. There are ways to work around that. All of them (and some more) were tried and/or are in use with IPv6: https://en.wikipedia.org/wiki/IPv6_transition_mechanisms https://en.wikipedia.org/wiki/IPv6_transition_mechanisms

rcxdude 5y ago
Those are all ways of allowing ipv6 to address ipv4, (or tunning ipv6 over ipv4 infrastructure) which is fairly easy to manage (getting the reverse connection still involves NAT though). None of them are ways of allowing ipv4 to address ipv6, because that fundamentally doesn't work by the pigeonhole principle.