On May 8, 2014, at 12:00 PM, <roe at liveblockauctions.com> wrote: I concur with others who are leery of the compute load. Some time ago, I blithely plopped a version of SSH on my NeXT 68040 cube: mtapley2> ssh -h ssh: SSH Secure Shell 3.2.9.1 (NEXTSTEP) on i386-next-openstep4 Copyright (c) 1995-2002 SSH Communications Security Corp SSH is a registered trademark and Secure Shell is a trademark of SSH Communications Security Corp (www.ssh.com). All rights reserved. See LICENSE file for usage and distribution terms. The comments are wrong, it?s a 68040, I?ve seen it, and the OS is NS 3.2. :-) Then, logging in from my MacBook on the same local net: mtapley3:dnetc518-macosx-amd64 mtapley$ ssh me at xxx.xxx.xxx.xxx <...17 clock seconds later...> me at xxx.xxx.xxx.xxx's password: (edited my user name and IP address, obviously). Here?s part of the report from ps -aux: mtapley2> ps -aux USER PID %CPU %MEM VSIZE RSIZE TT STAT TIME COMMAND mark 258 95.9 0.9 2.02M 608K p1 R N 4655hr ./dnetc -quiet ? root 173 0.0 0.5 1.38M 328K ? SW 0:06 (sshd2) ? Tolerable RAM footprint on the NeXT, looks severe to me for a PDP. Note, distributed.net is running in the background on the Cube and it?s not really as ?nice? as it claims it is, so an unloaded 68040 would be faster. However I think it?s reasonable to assume 10 seconds of crunch for a 25 MHz 68040 with nothing else running. sftp to that box is notably slower than ftp as well, If you need I can run comparisons but at this point it?s deep into YMMV territory. Hopefully that?ll help guess what a PDP 11 would be able to do. I have not done any attempt at optimizing or testing non-default encryption methods. - Mark