Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site brl-tgr.ARPA Path: utzoo!linus!decvax!genrad!wjh12!harvard!seismo!brl-tgr!internet!BJackson.PA@XEROX.ARPA From: BJackson.PA@XEROX.ARPA Newsgroups: net.unix-wizards Subject: Re: Need XNS on 3Com ethernet for 4.2bsd Message-ID: <5802@brl-tgr.ARPA> Date: Thu, 15-Nov-84 16:01:15 EST Article-I.D.: brl-tgr.5802 Posted: Thu Nov 15 16:01:15 1984 Date-Received: Sat, 17-Nov-84 05:34:08 EST Sender: news@brl-tgr.ARPA Organization: Ballistic Research Lab Lines: 36 Martin, I'd like present you with a few of my own personal observations that are based upon a wide ranging survey of existing "XNS" software. To start, I'd like to focus on you're statement: "I have 25 SGI Iris workstations talking XNS on an Ethernet." If you have Silicon Graphics Iris workstations, then they are using the low levels of the XNS protocol standards (mostly Sequenced Packet Protocol), but they currently do not and cannot coexist with Xerox workstations running the higher level XNS protocols. The least compilcated explaination of this is that the SGI software uses single net broadcast and name/address caching instead of an internet name service. SGI does have a software package which might help solve your problem, but my view is that this package DOES NOT CONFORM to Xerox NS principles due to its incomplete conformance to upper levels of the standards & lack of interneting capabilites. Chris Torek & James O'Toole and many others associated with the Xerox University Grants program are in the midst of trying to build an XNS package for 4.2 BSD which may provide the ability for 4.2 Vaxen to be full member on a Xerox NS Internet, but their solution will not be of much use to you because of the non-compilance of the Iris workstations. In other words, there are lots of hacks around that suggest that they do "XNS", but so far I have not seen any software that lives up to this expectation. I reccommend that anyone purchasing equipment which makes this claim look deeper than the sales literature at what's going on inside the software. Bill Jackson Computer Science Laboratory Xerox PARC These views are my own and do not necessarily reflect those of my employer.