kasper@csli.stanford.edu (Kasper Osterbye) (05/12/88)
Thanks to everyone who sent me information on migrating processes. I promissed a number of people that I would post my findings, and as the stream of messages have died down now, I think now is the time. I first list all the messages that I got containing references, and then I will list some that I have come across myself. ========= Marvin Theimer's thesis might be relevant to you. It's Stanford C.S. report STAN-CS-86-1128, "Preemptable Remote Execution Facilities for Loosely-Coupled Distributed Systems." -- Joe Weening Internet: weening@Gang-of-Four.Stanford.EDU Computer Science Dept. BITNET: weening%Gang-of-Four.Stanford.EDU@Stanford Stanford University UUCP: {decwrl,uunet}!Gang-of-Four.Stanford.EDU!weening ========= From: Peter Honeyman <honey@citi.umich.edu> the january 88 usenix had two papers on process migration. ousterhout's sprite also does migration; i don't know a reference offhand, but the most recen TOCS has a paper on sprite's file system, and that may give you a pointer. there is the classic xerox parc "worm" paper, which, again, i don't have a reference at hand for, but ask any old timer. peter >>I have the reference to the Worm programs at the end /Kasper<< ========= Fiction: VALENTINA - A Soul in Sapphire. Artif. Intel. program designed to be a game player grows to the point of self-awareness. EXCELENT story. Fiction: The Adolescence of P-1. Program designed to break system security and take unused resources for itself gradually grows to self-awareness. Good Story, but more along the Frankenstein tradition. Non-Fiction: The Tomorrow Makers. I've never read this myself, but it was mentioned on the network some days ago. Apparently it talks about up-and-coming people who are designing things for tomorrow. Including A.I. programs. ...!hadron\ "Who?... Me?... WHAT opinions?!?" | Edwin Wiles ...!sundc\ Schedule: (n.) An ever changing | NetExpress Comm., Inc. ...!pyrdc\ nightmare. | 1953 Gallows Rd. Suite 300 ...!uunet!netxcom!ewiles | Vienna, VA 22180 ========= You might want to look into RU (Remote Unix) developed at the University of Wisconsin. A paper was presented at the Summer 1987 USENIX Conference: Michael J. Litzkow, "Remote Unix: Turning Idle Workstations into Cycle Servers." pp. 381-384. Skip Montanaro (montanaro@sprite.steinmetz.ge.com, montanaro@ge-crd.arpa) ========= There's an article on "Fine-grained mobility in the Emerald System" by Eric Jul et al in the Feb 88 issue of ACM Trans. on Computer Systems. I think you may find it interesting ... If you need more information, feel free to contact me. - Rajendra K. Raj rkr@june.cs.washington.edu ========= From: jbn@glacier.stanford.edu (John B. Nagle ) The LOCUS distributed operating system, a distributed UNIX-compatible operating system developed at UCLA, supports program migration in a very thorough way. LOCUS offers a migrate(II) system call and a SIGMIGRATE signal. Active processes can be migrated to a different machine while running, without the process being aware that migration has occured. Open files, pipes, sockets, and signals are all handled appropriately. Migration only works between CPUs of the same type, of course; although LOCUS can support VAXen and SUNs, one cannot migrate between them. See "The LOCUS Distributed System Architecture", by Popek and Walker, ISBN 0-262-16102-8. ========= I wrote a survey paper "A Survey of Process Migration Mechanisms", that you may find useful. It will be in the next (July, I think) issue of ACM SIGOPS Operating Systems Review, or alternatively, I can mail you a copy (Send desired mailing address) or send one electronically, if you can print PostScript. It's chock full of useful references, and attempts to actually explain some of those things! -Jonathan jms@close.columbia.edu 450 Computer Science Columbia University New York, NY 10027 (212) 280-8181 (Jonathan M. Smith) ========= You can check out Marvin Theimer's PhD thesis, from Stanford, on process migration. I also have a paper in last year's Distributed Computing Systems conference ('7th ICDCS') and there was another paper in the most recent SOSP on the subject. There are plenty of other papers as well; look in these for references to some others. Hope this helps. Fred Douglis douglis@ginger.Berkeley.EDU ucbvax!douglis ========= From: OSCAR%CGEUGE51.BITNET@forsythe.stanford.edu Of course you have the reference to Knos in TOOIS. There's a new paper describing an implementation which appeared in COIS: %L Casa88 %K olit ooplas knos (geneva) %A E. Casais %T An Object-Oriented System Implementing KNOs %J Proceedings of the Conference on Office Information Systems (COIS) %C Palo Alto %D March 1988 %P 284-290 Cheers, Oscar >>One should considder never to use these ever so popular shorthands >>Not being in the field I have no idea of what TOOIS is /Kasper ========= Here at UCLA we are in the process of implementing Process Migration on an Intel iPSC hypercube. One reference " A Basic Protocol for Routing Messages to Migrating Processes" by T. M. Ravi and David Jefferson, 1988 International Conference on Parallel Processing, August 15-19, 1988. ravi ravi@cs.ucla.edu ========= And finally a few references of my own. John F. Shock, Jon A. Hupp. The Worm Programs - Early Experience with a Distributed Computation. Communications of the ACM 25(3) march 1982 The CHOROUS Distributed Operating System: Some Dessign Issues Marc Rozier, Jose Legatheaux Martins Institut National de Recherche en Informatique et en Automatique B.P. 105, 78153 Le Chesnay Cedex, France. Tel. (1) 39 63 52 94, Telex 697 033 F John Hogg Intelligent Message Systems Office Automation (Ed. D. Tsichritzis) D. Tsichritzis ObjectWorld Office Automation (Ed. D. Tsichritzis) I thought I had gotten a mad idea, but it seems like a lot of people is working on this. Hope that this will be to as much help to you as it was to me. Enjoy, Kasper kasper@csli.stanford.edu (internet)