On Tuesday 17 May 2011 15:44:39 Robert Hailey wrote: > > On 2011/05/13 (May), at 6:02 PM, Matthew Toseland wrote: > > > In any case, our options appear to be: > > > > 1) Try to make opennet work in China. > > ... > > > > We could try to rotate links even, so that only a few nodes have > > external connections at a time. The catch is that we don't know what > > the limit above is, and it will probably vary from time to time. So > > this is probably a dead-end. > > We would also need to be able to distinguish between a local & foreign > node. For a single case (like China) that could work, but the general > case would be rather difficult to code up (detecting censorship > boundaries: national or ISP...). Link latency *might* be clue enough > for national boundaries, but not blocking ISPs. > > > > > 2) Focus on darknet. > > ... > > Difficulties: > > a) If the Chinese darknet is completely sealed off from the western > > network, how would they even get software updates? We need better > > tools for migrating binary blobs. > > b) We need some way to ensure that FOAF connections don't result in > > dangerous external connections. > > Wether open or dark, if a connection crosses the firewall border it is > susceptible. If not dodging under the foreign bandwidth cap (as you > mentioned), then the only issue is what to do with the few (if any) > unfiltered connections. > > There is a chance that an algorithm which is both aware of 'separate > networks' and optimizes for latency would keep most of the traffic > within national borders (the firewall surely introduces some latency > by itself), and make it a non-issue.
There were some simulations of weakly connected darknets by vive. Apparently they mostly do work. > > One might predict the next iteration would block protocols not on a > whitelist. It can always get worse, but for internal comms not quickly IMHO. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part. URL: <https://emu.freenetproject.org/pipermail/devl/attachments/20110519/379a7e63/attachment.pgp>
