Dear all,

Following the discussion around draft-bartle-tls-deprecate-ffdhe, what are
your thoughts on deprecating RSA key exchange, and Finite-Field
Diffie-Hellman? (This would probably happen in a separate document.)

Considering the following different areas/use cases:
1. On the open Internet/web, both key exchange methods have been superseded
by ECDH. Browser support for FFDHE has been entirely removed IIUC, so
formally deprecating FFDHE should not be a problem (right?). Are there any
reasons to avoid deprecating FFDHE and RSA on the open Internet?
2. On local networks, deprecating both key exchange methods may leave some
endpoints without any key exchange algorithms. Could you please give
feedback on the following:
a. Is the number of such endpoints large enough that we shouldn’t deprecate
these methods?
b. If the answer to the above is yes, what would be a good plan/timeline to
deprecate them?

We could also consider limiting FFDHE to well-known groups of at least 2048
bits, with fully ephemeral secrets. But this would likely cause enough
interoperability problems that we might as well deprecate it fully, right?

thanks,
Nimrod
_______________________________________________
TLS mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tls

Reply via email to