ietf-nntp TLS cipher renegotation to NULL cipher

[Andrew Gierth]

>>>>> "Ken" == Ken Murchison <ken at oceana.com> writes:

 >> Andrew's hard data makes me think that it's more likely than not
 >> to be an issue, though.

 Ken> That might be true.  I'm curious how these numbers would change
 Ken> if the client (are we talking endusers or peers?)

endusers

 Ken> used session resumption and/or compression.

I did not address the cost of SSL session startup at all, only the
ongoing cost of keeping the session encrypted.

As I've repeatedly stated, there is no direct comparison (for me)
between the ongoing encryption cost and much of the cost of startup
(or all of the cost of a SASL mechanism like the one Chris Newman
proposed), because the latter can be farmed out to separate machines
and kept away from the high-bandwidth data paths. Session encryption
(or compression for that matter) on the other hand, increases the
overhead right where I can least afford it.

-- 
Andrew.