3 ms·
Is IPv4 shortage really going to be a problem? For client use, carrier-grade NATs allows to have 1000 IPs per customer, giving 6 million addresses. For server
by devit 7y ago
Is IPv4 shortage really going to be a problem?
For client use, carrier-grade NATs allows to have 1000 IPs per customer, giving 6 million addresses.
For server use, TLS SNI allows to have one IP per datacenter, which are estimated to be around 10 million in the world.
Non-TLS inbound usage is probably relatively rare, so overall around 100 million addresses should be fundamentally enough even accounting growth.
Of course there's a lot of inefficiencies, but the fundamentals seem to say that the IPv4 address space is enough.
- kstrauser 7y agoCarrier-grade NAT is a non-starter. It's fundamentally broken from an end user perspective and destined for the junk heap of forgotten bad ideas. TLS SNI is awesome, but no plausible implementation is going to let you put an entire datacenter worth of addresses behind a single address. That's not going to happen. IPv4's time is soon to be past. Google is nearly hitting 30% IPv6 usage already, and it's still growing. This is the way forward, not junk tech like CGN or millions of hosts on a single TLS SNI address.
- zokier 7y agoCGNAT (or comparable tech like DSlite) will happen anyways even when majority of traffic is IPv6. The long tail of IPv4 is long; people will be wanting some form of IPv4 connectivity long after all the major things have switched to IPv6 and ISPs are not going to go back to handing out public IPv4 addresses. The timeline just doesn't work anymore to make a clean transition. Of course how much do you care if it is CGNAT if you are not using IPv4 is another matter.
- zamadatix 7y agoI'd be willing to get there are more clients running through NAT64 than CGNAT currently thanks to mobile networks.