mason@PESCADERO.STANFORD.EDU (10/24/88)
> Comp.os.v is a newsgroup for discussions of the V operating system, >developed by David Cheriton's group at Stanford. It is not a widely used >system, even at Stanford, and is not likely to be; it's really a vehicle >for the development of new concepts in distributed processing. It runs >primarily on original Stanford-built Sun I systems, back-ended by VAXen. > > John Nagle Part of the above message is true, but part is not. For the record, I work in the Distributed Systems Group (headed by David Cheriton) and I can faithfully attest that we no longer *support* the Sun-I, let alone use them for new development. As for being "back-ended by VAXen" that isn't correct. We do have a VAX providing UNIX services, but most people within our group prefer using V here because it is so much faster. V currently runs on Sun-2's and Sun-3's, Microvaxen, DEC Fireflies, and our own in-house multiprocessor (VMP). It is in daily use by people outside our group, as well as those within it. As for not being "widely used" that is a choice that others make, not us. V provides the full range of services expected from an operating system, and also provides a true distributed computing base. We can hardly be held to blame if people wish to continue using inferior operating systems. As for the original message "delete comp.os.v and set up comp.os.mach," the fallacy here is there is some "conservation of groups." Since none such truly applies, the issues are separate. Establish comp.os.mach if there is demand, but don't trounce on our work. Tony Mason Distributed Systems Group Stanford University mason@pescadero.stanford.edu
jesson@killer.DALLAS.TX.US (John Jesson) (10/24/88)
In article <8810232014.AA16207@Pescadero> mason@PESCADERO.STANFORD.EDU writes: >> Comp.os.v is a newsgroup for discussions of the V operating system, >>developed by David Cheriton's group at Stanford. It is not a widely used >>system, even at Stanford, and is not likely to be; it's really a vehicle >>for the development of new concepts in distributed processing. It runs >>primarily on original Stanford-built Sun I systems, back-ended by VAXen. >> >> John Nagle > >Part of the above message is true, but part is not. For the record, I work >in the Distributed Systems Group (headed by David Cheriton) and I can >faithfully attest that we no longer *support* the Sun-I, let alone use them >for new development. As for being "back-ended by VAXen" that isn't correct. >We do have a VAX providing UNIX services, but most people within our group >prefer using V here because it is so much faster. > ...[other stuff deleted] Well, I'm very interested in distributed O.S... Do you think you could summarize about Vs aims and purposes in life? If an explanation would be too involved, how about recent references/T.R.'s concerning V? J.R. Jesson jesson@nrl-excalibur.arpa
brat@APPLE.COM (10/25/88)
> Where can we get a V that runs on Sun-2 and Sun-3? We still only have > the V6.0 dated back to '86 which is backed by Vaxen. I think you're confused (no offense intended). You are getting V confused with VMTP and Unix. V is an operating system, not a Unix server, and it has been running on Sun2's and Sun3's for many years (i.e., V6.0, the latest release of V, runs on these machines). What you're concerned about is whether or not you can have file service on a Sun with disks -- this is a different issue, and, yes, I agree that this ain't so easy in V6.0. The file server that runs in V didn't work then (can I claim it does now?), so you're forced to use a Unix filesystem, and to get that to work you have to have our Unix server running on a Unix machine. The Sun Unix version of our server was done by David Ungar here at Stanford but is not supported. I can't speak for Dr. Cheriton's intentions of cutting a new release tape anytime soon, but in the currently experimental version of V we rely very little on Unix services (still more than we should). But this is a research project and I'm mostly interested that people recognize the ongoing interesting research in the V system. We only use Unix because to build up a complete software base comparable to that of Unix in an experimental operating system takes alot of effort and time. There's no clear reason to commit the resources to doing that, but a phenomenal amount has already been done in that direction. When I was working on the V system, I worked entirely in V except where I needed to modify the Unix server and use MH. Steven