Hacker News new | past | comments | ask | show | jobs | submit login

IPv6 is a case study in the second sytem effect [1]. Realizing you need to make breaking changes and it being rare that you get to do so you decide to make all the changes.

The truth is IPv4 only had 2 real problems:

1. Lack of address space due to 32 bit addresses; and

2. Lack of a solution for roaming since your IP address is a core part of connection identity (between the source and destination address and port).

IPv6 managed to only solve one of these problems and it did so in a way that removed functionality. Yes there are more addresses but now address blocks are non-portable. I guess 25 years ago they thought that routing protocols like BGP4 were considered a problem and decided to remove that with hierarchical address spaces and massively expanded those address spaces to do it.

Somehow ports were also considered a problem so we got /64 addresses. Part of the motivation for this was to use (mostly) unique MAC addresses (48 bits) as your identifier and that fits in 64 bits. Of course this became a massive PII leak and a tracker's dream so it never happened but we're still stuck with /64 blocks that we absoultely do not need.

It's an object lesson in solving actual problems and not solving imagined problems.

[1]: https://en.wikipedia.org/wiki/Second-system_effect




Finally, a great comment in this thread.

Arguably lack of IP address block portability is not that big a deal [now that we're well used to it]. But IPv6 did not solve the CIDR route table size issues, and that's a big failure.

I have long thought that IP packets should have a source and destination ASNs as well as actual addresses. That would mean that we'd need more DNS (or similar) lookups, and a bootstrapping system for that. A scheme like this would greatly reduce router table sizes and would allow for IP address block portability.


You can get PI v6 space, and if you are somewhere where a RIR won't give it to you, that's not the protocol's fault.

Also ports weren't removed, they work just like in v4.

The long addresses with subnettable space for everyone is very valuable and useful and doesn't remove any functionality.


> Part of the motivation for this was to use (mostly) unique MAC addresses (48 bits) as your identifier and that fits in 64 bits. Of course this became a massive PII leak and a tracker's dream so it never happened but we're still stuck with /64 blocks that we absoultely do not need.

please read about rfc4941 privacy extensions & how prevalent their use is before continuing to regurgitate decade-old, outdated privacy alarmism about MaC aDdReSsEs.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: