3 ms·

That's a very interesting argument. It makes me that little bit more open to IPv6. But the standard argument is that IPv6 is detrimental to privacy because it

That's a very interesting argument. It makes me that little bit more open to IPv6. But the standard argument is that IPv6 is detrimental to privacy because it amplifies identity at the device level. So what's with the fixed IPs? The privacy upside to IPv6 is surely that, like mac address space, and rotating device IDs with macchanger, you can simply get lost in the vast IPv6 address space. Obviously I am equating privacy with anonymity here, but that's a reasonable part of a mature understanding of the "security, privacy, anonymity" triad. Are you sure the forces favouring centralisation have anything to do with the shortage of IPv4 address space? Would an IPv6 rich network really revive the original Internet design (peer-to-peer relations)?

Kim_Bruning 2y ago
I think it would! To be frank, even dynamic addresses don't provide that much extra privacy by themselves; but static addresses DO provide individual agency. Given individual agency, there's got to be hundreds of ways to secure your privacy. But without individual agency, your privacy will always be dependent on someone else. By having static addresses and permitting home services, you shift power away from centralized tech giants and bypass privacy issues caused by eg. content scanning, centralized storage and dependency on third-party systems. As just one example among many: right now it's very hard to provide IoT gear that "just works" without central services somewhere. If the equipment could just talk to each other directly; I do think that'd improve privacy by quite a bit! (among many other advantages)
- nonrandomstring 2y ago
  I think you make some good points. People got quite hung up on thinking of privacy at the IP/TCP layer because of legal disputes with RIAA/MPAA etc circa 2005 over IP addresses. It became lore that IP mobility/NAT/cloaking were the ways to protect yourself. But that's living on the run. With all the overlays these days, IPFS, Tor, wireguard tunnels, the game has moved to endpoints, so decentralising and creating peer diversity rather than network diversity is a win, as you say. But we should not entirely abandon ephemerality of network identity as a defence.

Kim_Bruning 2y ago

I think it would! To be frank, even dynamic addresses don't provide that much extra privacy by themselves; but static addresses DO provide individual agency. Given individual agency, there's got to be hundreds of ways to secure your privacy. But without individual agency, your privacy will always be dependent on someone else. By having static addresses and permitting home services, you shift power away from centralized tech giants and bypass privacy issues caused by eg. content scanning, centralized storage and dependency on third-party systems. As just one example among many: right now it's very hard to provide IoT gear that "just works" without central services somewhere. If the equipment could just talk to each other directly; I do think that'd improve privacy by quite a bit! (among many other advantages)

nonrandomstring 2y ago
I think you make some good points. People got quite hung up on thinking of privacy at the IP/TCP layer because of legal disputes with RIAA/MPAA etc circa 2005 over IP addresses. It became lore that IP mobility/NAT/cloaking were the ways to protect yourself. But that's living on the run. With all the overlays these days, IPFS, Tor, wireguard tunnels, the game has moved to endpoints, so decentralising and creating peer diversity rather than network diversity is a win, as you say. But we should not entirely abandon ephemerality of network identity as a defence.