CGNAT is stil NAT (and your solution will be same) - with reserved /10 just for this fun. Any kind of NAT breaks end-to-end connectivity, It's just a cripled internet, which relies on only one party establishing the connection. If you hide 200 users behind 10 IP addresses, it can't possibly turn out any other way. Yes, many end-users have already learned to live with it, or rather we "stupidly" teach them that it can't be done any other way. Again, both NAT and CIDR only came to delay the end of IPv4. And whole this discussion, hidden behind some ferocity, in reality it tries to do the same thing - just in a different way. But it was clearly stated 31 years ago that the only real solution is a new protocol. Even RIPE itself tried to delay the end - by gradually reducing the minimum allocation. Then politics came last /8 and with it a lot of speculators, as new LIRs were created just to grab the last /22. From short-term perspective, it was good, income was growing and there was no knowing what to do with the money. When the pond dried up, those speculative LIRs began to gradually dwindle and merge. That's cause of today's sustainability problem. understandably, this places increased demands on RIPE itself. Today's problems are primarily caused by speculation around those little "new" LIRs. And I don't blame NCC for this, from their position is this basically unmanageable (thanks to legal differences in region). And at the same time I think that the model based on IPv4 financing will not have that long a lifespan either. Maybe it will be in five years, maybe in ten - who knows. But one day you won't be able to get a visa over IPv4 and then things will start moving fast. And there're indications that some governments want to do this. And when this really happens, we'll be discussing here again how to fund RIPE, because IPv4 revenues will dry up... when we find out that there are actually no big differences between the members in terms of IPv6 allocations. It's the same story as with IPX protocol (for example) in LANs. It dominated her for a long time, but in the end it also died, forgotten. Even Linux removed it from it's kernel. A truly sustainable RIPE financing model must be independent on internet protocol version. If it's about real and long-term sustainability and not about something else motivated by short-term interests. - Daniel On 5/31/25 3:09 PM, sdy@a-n-t.ru wrote:
Pay attention to the CG NAT solutions.
I recently published such opensource code for linux, but it's in Russian language, so you'll need to work on it a little harder. Although someone cloned it without my consent in github, as if on my behalf.
Look for 'ANAT'. Or best go here:
NAT does not allow end to end connectivity. And if this discussion isn't about transitioning to IPv6, then the existence of NAT is not a valid argument either. IPv6 was created as a solution to the IPv4 shortage. NAT was created as a means to postpone the problem solved by IPv6.