[Top] [All Lists]

Re: [ietf-smtp] guidance on how to secure against sniffing and paid backdoors

2013-09-14 09:41:55
On Sat, 14 Sep 2013, keld wrote:
Recently there has been reports in newspapers about powerful organisations that can sniff on wires and has paid for backdoors and compromising cryptographic implementations.

Would it be a good idea to make a document describing best practices trying to protect against such actions, to guide implemetors and service providers?

If "such actions" involve backdoors in implementations and standards, and possibly secret mathematical knowledge that can help crack cryptography, then protecting against them is beyond the scope of email. (In this particular case, it's probably beyond the technical realm.)

A separate question, that others are trying answer in this thread, is: assuming we have working cryptography, how can we use that to make email more secure? That's a tough question and I think that ultimately, if information is to be kept really secret, it shouldn't be sent over email.

In cases where the fact that communication takes place is not a secret, and you're not worried about metadata leaking, something like PGP would probably work. But that doesn't scale. And in many cases metadata does matter.

Of course, the fact that certain agencies can, in principle, read all your emails shouldn't stop you from protecting it from the prying eyes of others. So using SSL/TLS wherever possible is a good thing. But a few commercial providers offering what's been standard in many other systems for years, like the "email made in Germany" project, is a marketing stunt at best.


ietf-smtp mailing list