Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!ucbvax!cayman.com!josh From: josh@cayman.com (Josh Littlefield) Newsgroups: comp.protocols.appletalk Subject: Re: Liaison, AppleShare, and other fun stuff Message-ID: <9001122032.AA16702@cuba.Cayman.COM> Date: 12 Jan 90 20:32:28 GMT References: <1990Jan12.080330.935@caen.engin.umich.edu> Sender: daemon@ucbvax.BERKELEY.EDU Organization: The Internet Lines: 33 In artical <1990Jan12.080330.935@caen.engin.umich.edu> you writes: >> In article <9001111621.AA03743@icarus.cns.syr.EDU> demarsee@ICARUS.CNS.SYR.EDU (Darryl E. Marsee) writes: >> >A few questions: >> > >> >1) I'm trying to connect a Mac over a modem connection to our campus >> > AppleTalk using Liaison 1.0.1. .... >> This is because atalkad is a static system. A good rule of thumb is that >> you cannot connect any router to an atalkad network unless it is a KFPS >> running K-STAR (or KIP), and is entered into the atalkatab for atalkad. >> >> Here is why. First fact: K-boxes do not send RTMP packets along the routes >> they receive via atalkad. ... This is not completely true. If you make one or some of your FastPaths a core router (add the "C" flag next to the "K" flag), then all the atalkad FastPaths will exchange routing information via the "AA" protocol (not exactly RTMP). The basic idea is that once per minute each non-core router sends an aaROUTEQ packet to one core router, containing locally discovered routes. The core router responds with an aaROUTE packet containing all routes it knows about which it didn't learn from the atalkad daemon. Since the routing tuples are "3rd party" the core router does not become a bottleneck to traffic, only a clearinghouse for routes. -josh ------------------------------------------------------------------------------ Josh Littlefield Cayman Systems, Inc. University Park at MIT josh@cayman.com 26 Landsdowne Street (617) 494-1999 Cambridge, MA 02139 ------------------------------------------------------------------------------