{ "comments" : [{"author": "Michael", "email": "135b6fda84815053ef52ef72fe54927d", "website": "", "submitted": "2013-01-07 22:18:54", "comment": "

Two comments: one, the links in the table 404 (they link to ritter.vg, not crypto.is)

\n

Second, as a request for comment / future blog post subjects: what strategy does the recipient have for authenticating a received message? If john@nytimes.com gets an remailed message from a user claiming to be, let's say, a whistleblower exposing corruption in the Fortune 500 he works for -- how does john verify that the sender is in fact an employee?

"},{"author": "Tom", "email": "31e32ecb7309ad47e1ecdd34f4c26529", "website": "http://ritter.vg", "submitted": "2013-01-07 22:34:32", "comment": "

Thanks, fixed the links. I'm not sure that would be a good thing to cover (except as a \"nota bene\") because the short answer is there is none. Remailers provide absolutely no sender information by design, so if someone said they were a Fortune 500 whistleblower, all authentication would have to happen either out-of-band (e.g. via a meeting) or by some way the reporter would 'trust' - providing information that would convince anyone of the whistleblower's status. There's no technical measure remailers provide that would convince anyone of anything.

\n

Furthermore, you'd need to make sure that the person you were receiving information from via the remailer was always the same person which each new email! If I knew you were communicating with a whistleblower, I could send you an anonymous email pretending to be the whistleblower saying \"Something has happened, call me immediately!\" and then unmask the whistleblower by seeing who the reporter calls!

\n

Contrast that with nymservers. Nymservers provide a linked cryptographic identity. So if I get an email from Alice at a nymserver, and then a day later again from Alice at a nymserver - I can trust that this is the same person up to the point we can trust any email not to be forged. Which is not a whole lot really. But the confidence does go up if there's a valid SPF and DKIM. But again, with nymservers, you still need to make sure Alice really is a whistleblower by some outside means.

\n

So... now that you mention it, putting all this into a blog post on Operational Security would be a good idea. =P

"},{"author": "Michael", "email": "135b6fda84815053ef52ef72fe54927d", "website": "", "submitted": "2013-01-11 22:49:55", "comment": "

I wouldn't call it Operational Security -- I'd call it \"practical matters\" -- though maybe that's the same thing. For example, you can't use PGP because that's tied to the sender identity. And you probably shouldn't say \"meet me at the starbucks at 5th and Main at noon Thursday\". What can he say? How do you prove your identity without proving your identity?

\n

If john@nytimes.com, via some OOB channel, knows the true sender address, should he respond via a remailer? Is receiving a remailed-message considered suspicious by network observers?

\n

Yeah, maybe \"Operational Security\" and \"practical matters\" are the same thing.

"},{"author": "Niel.S", "email": "282d92f23960333b95c9427a5eb53e45", "website": "http://link.springer.com/chapter/10.1007%2F11927587_25", "submitted": "2013-04-25 14:50:08", "comment": "

A third solution to anonymous channels, Klein bottle routing, is proposed in this paper. It fills the gap between onion routing and mix network and can be widely .

"},{"author": "Michael", "email": "135b6fda84815053ef52ef72fe54927d", "website": "", "submitted": "2013-04-27 16:24:35", "comment": "

Another question: what advantages/disadvantages does a remailer present over using TOR to connect to a single-purpose hotmail account?

"},{"author": "Tom", "email": "31e32ecb7309ad47e1ecdd34f4c26529", "website": "http://ritter.vg", "submitted": "2013-05-01 14:59:23", "comment": "

Well, that's the rub right? Tor+Hotmail has obvious advantages (usability) and disadvantages (a hostile party managing your communication).

\n

Remailers have a very weird reply or conversation model. I don't know if I would trust, or anyone has examined, how much data is leaked by a normal thread of conversation between people using remailers. A standard email account's is normal. In some scenarios Mix Networks are more theoretically secure than Onion Routing. In general, Tor may be more secure than remailers. Introducing a third party (hotmail) who should be assumed to be hostile to you is sketchy.

\n

This is a question I debate in my head regularly.

"}] }