[throwaway32]

The benefit is real end to end connectivity without the need for co-operation of everyone along the chain (this is an extremely serious issue, often even one layer of NAT makes trying to get port forwarding a nightmare, imagine 2 or 3) . Its also much more scalable than both. In the long run supporting some kind of gargantuan NAT/SOCKS translation layer for millions of customers is going to be much more expensive than ipv6 support, because IP is fundamentally stateless and designed to scale massively, NAT and SOCKS are not. This becomes a serious concern at larger scales.

[alyrik]

SOCKS gets you end to end (though it's arguably not "real"!) by letting you have listening sockets (which NAT doesn't).

There is indeed a scaling issue with NAT/SOCKS as both need connection state on the gateway device - but, I suspect, the dropping cost of hardware versus the rising cost of IPv4s will favour more hardware. The real question is how it'll stack up against the cost of dropping IPv4 for IPv6 (not the cost of adding IPv6 support alongside IPv4, which isn't that hard).