There's no chance in hell class E, 127/8 or even 0/8 will work within 20 years.
The effort is doomed.
If you thought it was a lot of work to upgrade to IPv6, forget about it.
- Hey ISP or cloud provider, do you want some of that sweet sweet class E?
- Will users be able to connect to it?
- Nah. Well, some will. As long as you're happy with people spending the next 20 years complaining about how "your servers are always broken", you can have some addresses.
It's a complete waste of effort that should instead be spent adding support for IPv6.
Spend huge effort on something temporary throw-away, or on something long term? The temporary thing will take at least as long to implement, and you will never know if you're actually done or not.
The effort is doomed.
If you thought it was a lot of work to upgrade to IPv6, forget about it.
- Hey ISP or cloud provider, do you want some of that sweet sweet class E? - Will users be able to connect to it? - Nah. Well, some will. As long as you're happy with people spending the next 20 years complaining about how "your servers are always broken", you can have some addresses.
It's a complete waste of effort that should instead be spent adding support for IPv6.
Spend huge effort on something temporary throw-away, or on something long term? The temporary thing will take at least as long to implement, and you will never know if you're actually done or not.