At 12:46 PM -0800 4/22/98, Jonathan S. Shapiro wrote:
>This prompts me to two thoughts:
>
>1. I had not previously appreciated the degree to which EROS/KeyKOS
>applications are able to run out of "virgin" frames and therefore
>avoid the overhead of faulting them in. This should have significant
>(positive) impact on performance that I had not adequately understood.
>
>2. It is a surprise to me just how small applications on this system
>are proving to be. While the kernel is a lot bigger than I might have
>liked, the application-level stuff is proving to be pretty tight.
>Jonathan Adams recently made a similar comment.
We noticed this as well when we used KeyKOS to do terminal protocol conversion. The resulting application connects a X.25 public network to a mainframe and uses common ASCII terminals (e.g. VT100) to simulate the native (proprietary) full screen (character mode) terminals. It will support at least 100 simultaneous users in 16 Meg of main memory. (It does not use the disk at all.)
Bill Frantz | If hate must be my prison | Periwinkle -- Consulting (408)356-8506 | lock, then love must be | 16345 Englewood Ave. frantz@netcom.com | the key. - Phil Ochs | Los Gatos, CA 95032, USA