Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B UNSW 1.1 19 Sep 1984; site elecvax.OZ Path: utzoo!watmath!clyde!burl!ulysses!allegra!bellcore!decvax!genrad!panda!talcott!harvard!seismo!munnari!basser!elecvax!michaelr From: michaelr@elecvax.OZ (Michael Rourke) Newsgroups: net.lan,net.unix-wizards Subject: XNS on UNIX (Summary of responses) Message-ID: <1065@elecvax.OZ> Date: Sun, 19-May-85 21:17:56 EDT Article-I.D.: elecvax.1065 Posted: Sun May 19 21:17:56 1985 Date-Received: Thu, 23-May-85 03:28:16 EDT Organization: EE and CS, Uni of NSW, Sydney, Australia Lines: 141 Xref: watmath net.lan:810 net.unix-wizards:13271 In article <886@elecvax.OZ> I wrote: We are interested in getting hold of some XNS source to run on our UNIX system (we have System V and bsd4.2 licences). Has anybody written anything in the public domain? I would also be interested to hear from anybody with experiences (good or bad) with this network protocol under UNIX. Interlan offers the source for $85K which seems a bit over the top! (we could write it for much less than that). PLEASE REPLY BY MAIL - My host isn't receiving netnews. Here are the responses received (edited of extraneous headers): -------------------------------------------------------------------------- From: Mark Weiser We at maryland have put XNS up through spp into the 4.2 kernel. We reported on this at the Dallas Usenix in Jan. Cornell has put Courier on top of our spp. We talk regularily to our network of Dandelions running STAR and XDE using this stuff. We charge $250 for our software distribution tape (which includes lots of other stuff). Send mail to seismo!umcp-cs!despina to get the proper order forms and more information. Spoken: Mark Weiser ARPA: mark@maryland Phone: +1-301-454-7817 CSNet: mark@umcp-cs UUCP: {seismo,allegra}!umcp-cs!mark USPS: Computer Science Dept., University of Maryland, College Park, MD 20742 -------------------------------------------------------------------------- From: topaz!hedrick@seismo.UUCP (Chuck Hedrick) Organization: Rutgers Univ., New Brunswick, N.J. If you get any responses, make sure that you know what they mean by XNS. A number of implementations include only the packet transfer protocol. If you just want to talk with another implementation that uses that level, fine. But if you want to talk with a Xerox system of one sort or another that uses Courier, then that won't solve your problem. We have yet to find a non-Xerox implementation of Courier. Cornell is working on one to run on top of the Univ. of Maryland XNS implementation. -------------------------------------------------------------------------- From: Chris Torek We have XNS code for 4.2BSD that we are distributing. We have implemented all of IDP and SPP, and written a few Mesa utilities that pretty much act like telnet and FTP. We speak to Xerox Dandelions, which indicates that the code might indeed be working, as well as to our own Vaxes. Sorry, no performance data yet.... we're working on that now, though. Send mail to Diane Miller Computer Science Department University of Maryland College Park, MD 20742 USA for more information. -------------------------------------------------------------------------- From: unido!ecrcvax!pete@mcvax.UUCP (Pete Delaney) Here in europe we are establishing an ISO implementation that could become PUBLIC domain, at least here in europe. I suggest you talk to mvcax!vmucnam!lvbull!jim about the ROSE project under the European Commission (ESPIRT). -------------------------------------------------------------------------- From: ut-sally!oakhill!mot!fred@seismo.UUCP why not try Network Research Corp, Santa Monica, CA? 213-394-7200, as i recall. that's what we run. Fred Christiansen, Motorola Microsystems, Tempe {ihnp4,allegra}!sftig!mot!fred {ihnp4,seismo}!ut-sally!oakhill!mot!fred {ihnp4,amdahl}!drivax!mot!fred -------------------------------------------------------------------------- From: rochester!rocksvax!z@seismo.UUCP (Jim Ziobro) The XNS protocols are being worked on at several places. The University of Maryland has done the kernal work and Cornell is doing a Courier implementation. I hope to get more details at the upcoming USENIX. The protocols themselves work very well. I expect within the year all of Xerox will be using them. -------------------------------------------------------------------------- From: harvard!enmasse!enms2!soph!dave@seismo.UUCP Organization: EnMasse Computer Corp., Acton, MA As well as Interlan, there's also Network Research corporation somewhere in California. Phone -- (213)394-7200, I think. I have an address that may be good: 1101 Colorado Ave, Santa Monica CA 90401 USA. Their FUSION package handles XNS and TCP, and they do offer source licences. We have been rather displeased with them, however the software does work if you don't push it TOO hard. It runs under many systems (including Sys V and BSD 4.2). They are not a very professional outfit. Re experiences: the protocol family itself seems fine. The Stream protocol would seem to have flow control problems in networks with fair sized delays (e.g. satellites), but it's certainly smaller/faster than the DoD family. Dave Brownell enmasse!dave@Harvard.ARPA {genrad,harvard}!enmasse!dave -------------------------------------------------------------------------- From: sdcsvax!sdchema!jwp@decvax.UUCP@mulga (John Pierce) Organization: Chemistry Dept, UC San Diego Sorry I can't help you with code, but we do have some experience with running XNS... We had a Silicon Graphics graphics terminal that had to talk to one of our VAXen with XNS (all the SG machine supported). It was an unpleasant experience... Couldn't use the ARP code, which forced us to "fake" net numbers until we could upgrade the terminal to a full workstation (which is supposed to support tcp/ip, but doesn't work right for some reason). The experience of another group on campus that has had to use XNS was the same as ours - no way to get it to work and have ARP running at the same time to talk to machines on other networks (the stuff using XNS was supposed to be on a separate network). This may be due to buggy code in 4.2, of course, since it looks like it *should* work but I've got a feeling that it really has to do with something inherent in the way the XNS stuff has to work. John Pierce, Chemistry, UC San Diego -------------------------------------------------------------------------- From: enea!kuling!hans@mcvax.UUCP (Hans Eriksson) Organization: CoSyL, Upsala University, Sweden We are also on the search for Unix/XNS software. We have found something, we think, at Univ of Rochester but we do not yet know what it is. I would appriciate forwarding of any info you get on this subject. I will tell you what we found in Rochester, when we know. Hans Eriksson -------------------------------------------------------------------------- As before, could any further responses be mailed to me, as we still aren't receiving net news properly. Michael Rourke School of Elec Eng & Comp Sci ACSNET: michaelr@elecvax.oz University of New South Wales UUCP: {decvax,vax135}!mulga!elecvax.oz!michaelr Kensington NSW 2033 AUSTRALIA Phone: +61 2 697 4040