> On Dec 1, 2020, at 4:26 PM, Phillip Hallam-Baker <phill@xxxxxxxxxxxxxxx> wrote: > > The gap between the theory as you understand it and implementation at the time was huge. The gap between your implementation and our measured experiments on commercial versions appears to be bigger. We saw connection limits at ftp.isi.edu but that was for socket exhaustion per service per client, in the thousands per 2 minute time wait. Not 8 - and we never override the defaults. Joe
