On Sun, 27 Oct 2013 20:17:30 -0400 Bill Cox <waywardgeek@xxxxxxxxx> wrote: > Thanks for the links, Roger. On the first thread, I agree with you > that exit nodes should not be expected to analyze the unencrypted > traffic to determine if behavior is acceptable. That's what NSA exit > nodes are for :-p However, OpenDNS style filtering should be > acceptable, IMO. It'd be a simple set of flags that each exit node > says it supports or blocks, so it could be fairly simple for Tor > users to route to an acceptable exit node. I would want the same > flags for relay nodes as well. That would block some sites people > don't want to redirect, such as porn sites > - not that I have issues with porn. I just don't want to waste my > bandwidth routing it. I'm really confused how you expect "middle relays" to implement this sort of filtering, given that *by design*, the middle relays are ignorant of the destination of the traffic. Oh I know, we could add a bit in the cell header that signifies if the cell is carrying "objectionable or evil" payload a la (RFC 3514), and it will be up for compliant implementations to tag their packets when they're watching pron, so that relays can filter/censor traffic as appropriate. There's probably fairly "interesting" anonymity implications when certain relays won't ever be eligible to be part of your path as well. (As a side note how do you know that they were watching pron through your exit? It's usually considered a bad idea for various reasons to look at user traffic.) Regards, -- Yawning Angel
Attachment:
signature.asc
Description: PGP signature
-- tor-talk mailing list - tor-talk@xxxxxxxxxxxxxxxxxxxx To unsubscribe or change other settings go to https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-talk