On Fri, Sep 19, 2014 at 1:38 PM, Tim Bray <[email protected]> wrote: > A number of things about this one made me kind of uneasy. The clichéd tone > of the article “high-school dropout trumps NSA!” The complete absence of > input from anyone who wasn’t a project insider, and the dissing of > competitors who are actually shipping working software.
Seems like a critique of the journalism, not the project. The project looks like a simple-ish chat protocol using Tor Hidden Services for peer-to-peer connections. I think it relies on Tor HS for encryption and server-auth, and adds some fairly simple client-auth. There's a slew of new apps using the model of "Tor Hidden Services for peer-to-peer connections". We had a thread about that for email-like messaging, e.g. https://moderncrypto.org/mail-archive/messaging/2014/000434.html https://moderncrypto.org/mail-archive/messaging/2014/000447.html The advantage of this model is that your metadata isn't seen by a server. A less obvious disadvantage is that, compared to proposals where both parties use Tor as clients to communicate via some server, users might be exposed to things like: - hacking / DoS targeted at your Hidden Service - deanonymizing users via Hidden Service attacks - deanonymizing users via monitoring HS uptimes - linking users via monitoring Alice's HS uptime, and correlating it with Bob's polling to see if Alice is up I also don't know how well Tor HS would scale to large numbers of people using it this way. But that would be a good question for a HS expert (do we have one?) Trevor _______________________________________________ Messaging mailing list [email protected] https://moderncrypto.org/mailman/listinfo/messaging
