Hacker News Reader

HNHacker News

Top New Best Ask Show Jobs

Comment by blueflow | Hacker News Reader

Parent Full thread

blueflow·

NAT alone is definitely worse than IPv6.

View on HN

[Citation needed]

We've been using NAT and expanding it to decades which makes it an easy proof that it brings value.

IPv6 is still barely adopted and barely available in two decades after having "ipv6 parties". What does that tell you about its value?

40%-45% of Google users use IPv6[1]. I wouldn't call that "barely adopted".

[1] https://www.google.com/intl/de/ipv6/statistics.html

true, though i would guess that's mostly phones and it seems to be stagnating

APNIC stats show that assigned IPv6 /32 prefixes are still growing YoY. Wouldn't call that stagnating.

https://rex.apnic.net/statistics?allocationType=asn,ipv4,ipv...

was referring to the google stats, which seems to be approaching 50% v6

And those phones are behind an equivalence of CGNAT so the whole dream of "no NAT" is utterly not true for most IPv6 deployments.

Only for IPv4 destinations however, where there is no other way. For IPv6 destinations it's just native connectivity with no NAT.

Through a stateful firewall which blocks all incoming connections, completely invalidating "every device accessible" tenet of IPv6 and bringin zero value to the actual customers.

NAT has state that needs to be tracked for packages to be routed correctly. The state memory is limited, so at some point it will be dropped and you will get the famous "connection reset by peer" which has been plagueing the internet for the last decades.

Do the same with IPv6, omit the NAT, and you do not need to keep that state in sync. P2P applications stop being a routing/port forwarding problem and are only a matter of firewall settings. Cut away all that SIP and IPsec debugging.

Reply on news.ycombinator.com