3 ms·

Any pipe to the US would have challenges on the northern end as well. The connection has to connect to something, after all, and the US government could just a

by bathat 14y ago

Any pipe to the US would have challenges on the northern end as well. The connection has to connect to something, after all, and the US government could just as easily apply pressure at the point of presence to any backbone provider who peers with "MegaPipe". It seems like the idea here is to prevent Mega's "legally challenged" (as another commenter put it) servers from being disconnected by going all the way up the 'chain of command' and creating his own backbone. He intends to give the backbone legitimacy (to forestall any funny business at the PoP) by providing free internet to New Zealanders, thus generating large amounts of 'legitimate' traffic on the link. Edit: Just to be clear, like rdl, I don't actually think this is a worthwhile thing to do.

rdl 14y ago
There are some international treaties about submarine cables (mainly voice termination) which make it somewhat better than "they'll just unplug the pipe at the other end"). Mainly a cable system which goes from NZ to US might have to carry traffic from NZ to US to MX even if it blocks NZ to US traffic. This may extend to other traffic. (some of this is human rights related, and I mainly know the EU/UK versions since that was what was relevant to me, and pre-9/11. The US is a special case of a country which both can and does do essentially whatever it wants. Running your circuits to a bunch of other countries would make more sense.) You could also build a cable system with huge bandwidth, 99% legit NZ origin traffic (or onward transport from AU), encrypt everything, run the cable to Los Angeles, and then buy a bunch of transport from 1 Wilshire to various locations outside the US. Encrypt your "special" traffic all the way to the various locations, announce your routes there. None of it is airtight or bulletproof, but you can add layers of obfuscation, some of which also slow legal proceedings (some of which are still civil, some are criminal). Still probably not worthwhile. Ultimately the best strategy for "challenging hosting" is to replicate the content itself as widely as possible, then build some kind of blacknet to send transactions and command and control. This is capable of resisting arbitrarily strong efforts at censorship. (A great way is to distribute encrypted data widely, then drop the key at some point in the future. That's the "wikileaks insurance file" approach.)

by bathat 14y ago

rdl 14y ago

There are some international treaties about submarine cables (mainly voice termination) which make it somewhat better than "they'll just unplug the pipe at the other end"). Mainly a cable system which goes from NZ to US might have to carry traffic from NZ to US to MX even if it blocks NZ to US traffic. This may extend to other traffic. (some of this is human rights related, and I mainly know the EU/UK versions since that was what was relevant to me, and pre-9/11. The US is a special case of a country which both can and does do essentially whatever it wants. Running your circuits to a bunch of other countries would make more sense.) You could also build a cable system with huge bandwidth, 99% legit NZ origin traffic (or onward transport from AU), encrypt everything, run the cable to Los Angeles, and then buy a bunch of transport from 1 Wilshire to various locations outside the US. Encrypt your "special" traffic all the way to the various locations, announce your routes there. None of it is airtight or bulletproof, but you can add layers of obfuscation, some of which also slow legal proceedings (some of which are still civil, some are criminal). Still probably not worthwhile. Ultimately the best strategy for "challenging hosting" is to replicate the content itself as widely as possible, then build some kind of blacknet to send transactions and command and control. This is capable of resisting arbitrarily strong efforts at censorship. (A great way is to distribute encrypted data widely, then drop the key at some point in the future. That's the "wikileaks insurance file" approach.)