3 ms·
"Just extend the address size" was certainly one of the options. Sure, it's still a change, but the point is: After this change, both protocols could have worke
by bradley13 3y ago
"Just extend the address size" was certainly one of the options. Sure, it's still a change, but the point is: After this change, both protocols could have worked side-by-side. Devices that only supported IPv4, no problem, they send 32-bits. Devices that supported IPv6-as-it-could-have been would simply have zero-padded those 32-bits to match the new protocol. Talking to old devices, the zero-padding gets dropped.
That would have saved a lot of pain.
- lmm 3y agoNo it wouldn't. Think about how that would work for a minute. It would've created nondeterministic routing loops that would be a nightmare to debug.
- josephg 3y agoThen any network address beyond ipv4’s 32 bit range would have been completely inaccessible to any legacy devices. That would have essentially been the same situation that we have now - where ipv6 only services are inaccessible to anyone on an ipv4 network. So service operators need to keep their ipv4 addresses and networks don’t update. How would that be an improvement over the existing situation?