8 ms·
If you do layer 4 network address translation (NAT) - then a single IPv4 address can theoretically be used by circa 64000 people at the same time; depending on
by ay 15y ago
If you do layer 4 network address translation (NAT) - then a single IPv4 address can theoretically be used by circa 64000 people at the same time; depending on how multiplexing is done, by maybe even more.
The large-scale translators on the headends are the reality for quite a few service providers already.
So, the "IP address does not correspond to an individual" is not just a mere viewpoint gaining momentum, but an emerging hard technical fact. ;-)
- Jach 15y agoTinfoil hat Just wait, once the US government at large finally understands IPv4 networking, someone with too much power will look at IPv6 and the phrase "Look we don't have to do NAT anymore!" Then the government will assign a unique IPv6 address or two to every citizen to use as ID (they'll need a replacement anyway once they get rid of social security). Then they mandate the ISPs to track IPs, and they make it illegal to share or mask IP/IDs since that's something only pirates and thieves and other bad people do anyway. And we'll look back and say "Remember The Right to Read? rms was right again!"
- kiloaper 15y agoUnique internet IDs are far from a conspiracy theory. They're being actively developed in China and South Korea to name a few. While the South Korean one has faced setbacks it's only a matter of time before they become common place in other countries in my opinion.
- ay 15y agoIf governments were to do that, they would also be forced to solve the problem of the routing table explosion. Which is a hard problem. As for unique ID - it's already doable and being done without any IPv6. Read up on evercookie. There was an article on HN a while ago about a bunch of advertisers doing just that. This is worse than the government having it - not only it gets abused, but someone gets rich on this.
- tedunangst 15y agoThere are 16 million IPs in 10.x.x.x.
- fryguy 15y agoOnly 16 bits worth of ports, though.
- premchai21 15y agoBut there are at most 2^16 TCP ports available on the other side. I think this doesn't theoretically prevent one from reusing an exterior port multiple times for multiple connections to different other IP addresses or ports, but that would be both rather tricky and inconsistent with some cultural expectations regarding the use of initiating ("client") ports, and so I doubt any current NAT stack in common use uses that approach (though I'd love to be proven wrong). It also fails naturally if all sixteen million hosts try to communicate with the same host/port on the other side. UDP has similar but nonidentical constraints.
- eggnet 15y agoOnly this tuple needs to be unique: (source ip, source port, dest ip, dest port, protocol) For protocols that don't have ports, then it is this tuple: (source ip, dest ip, protocol) that needs to be unique. Openbsd's pf works that way, I'd be utterly shocked if any other linux or bsd kernel-integrated nat system worked any differently.
- premchai21 15y agoI was originally thinking this would break things like IDENT, but then I remembered that they use the entire tuple (d'oh). I'll take your word for the pf behavior, and retract my bogus guess about kernel NAT. :-)
- tedunangst 15y agoOh, good point, I didn't really think of "at the same time" as actually meaning simultaneous connections to the same dest.