On 2022-04-05 15:00, Philip Balister wrote:
Does the Ettus image have iperf3 on it? I suppose I should remember
how it works and see if it has useful numbers.
That would help find the bottle neck.
Philip
It's almost certainly the encryption, given that there's a 3.5:1
improvement going from "ssh" to "in the clear over straight HTTP".
Should probably try running SSH with all the verbosity turned on to see
which algorithm it's negotiating. If it's anything other than
AES I'd be surprised, but other algorithms, for example, the 3DES are
quite slow (relative to AES) on computer hardware, due to the
arbitrary permute that is used between rounds. That's not easy to
speed-up in software.
AES, on the other hand, is "modern", and was designed to be "speedy" on
a wide variety of computer hardware. I wouldn't bet that
ARM-based AES implementations win any performance prizes...
/me former crypto weeny
_______________________________________________
USRP-users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
