Re: Telnet Encryption Specification

[John Hawkinson <jhawk at mit.edu>]

| Just as a point of reference, I regularly use ssh (which has quite a
| bit of overhead) for interactive traffic over a modem, and I can't
| detect any difference in latency over a non-protected connection.  I
| think the days where we need to worry about the extra overhead are
| also over.

Another important consideration (the first being the CPU issues) is
behavior in the face of suboptimalities in network performance
(i.e. loss or excessive jitter). Yesterday I was fighting a nasty
battle against both and found ssh to be unusuable for interactive
work, even though telnet worked acceptably.

I'm not sure if the CPU commentary is meaningful in a CBC vs. OFB/CFB
discussion, versus an "encrypt vs. not encrypt" discussion.

--jhawk