CURRENT_MEETING_REPORT_ Reported by Steve Deering/Xerox Parc, from notes by Jeff Mogul and James VanBokkelen MINUTES This was the second meeting of the Router Discovery (nee Gateway Discovery) working group. Jeff Mogul served as acting chair, in Deering's absence. The proposed protocol from the December meeting was reviewed. The significant features are: o It is an ICMP extension. o Routers periodically multicast router reports; hosts multicast router queries at boot time only. o Use of broadcast instead of multicast is permitted but discouraged. o A router report does not include a subnet field. o A router report includes a holding-time field and a preference-level field. o In cases where more than one subnet is assigned to the same physical network, a router may include multiple addresses (i.e., one for each subnet) in a single router report. Jeff identified the following open issues: 1. Meaning of preference levels: Nobody seemed to know what to do with more than three levels (primary, backup and last chance?). Decision: use 8 or 16 bits, whatever fits in the packet format. 2. Choice of multicast period: Noted that ES-IS uses 10 seconds; we may want slower? 3. How should router respond to query, unicast or multicast? Mike Karels proposed that routers be allowed to attempt to avoid unnecessary replies, substituting a single broadcast or multicast for several unicast replies. Decision: "keep it simple", i.e., always send unicast replies. 4. Who writes the RFC? No volunteers, so it's still Deering's responsibility. 5. Do clients dally before sending queries? Yes, so that if a LANful of X terminals reboot from ROM at the same instant, they don't all hit the broadcast at the same instant. Other issues raised: o Use on non-broadcast media. Dismissed as too complicated. o Do routers dally before replying? Someone suggested that the router dally randomly (mean time based on pref level) to avoid overwhelming client. We argued over 1 whether the clients needed all the possible router responses right off. However, since we don't want to invent a protocol to stop the other N routers from responding, we realized that if we were already going to expend the resources for N+1 packets on the wire, we might as well try to make use of them at the client. So, dallying seems to be wanted. o When to query? Drew Perkins proposed a minor shift of definitions to allow initial query to be sent when a gateway is first needed (i.e., when first sending to an off-subnet destination), rather than at boot time. ATTENDEES Ballard Bare bare%hprnd@hplabs.hp.com Art Berggreen art@sage.acc.com Richard Bosch probe@mit.edu Ron Broersma ron@nosc.mil John Cavanaugh John.Cavanaugh@StPaul.ncr.com James R. Davin jrd@ptt.lcs.mit.edu Farokh Deboo fjd@interlink.com Rich Fox sytek!rfox@sun.com Mike Karels karels@berkeley.edu Tony Mason mason@transarc.com Keith McCloghrie sytek!kzm@hplabs.hp.com Bill Melohn melohn@sun.com Jeff Mogul mogul@decwrl.dec.com John Moy jmoy@proteon.com Drew Perkins ddp@andrew.cmu.edu Michael Petry petry@trantor.umd.edu Mark Rosenstein mar@mit.edu Tim Seaver tas@mcnc.org Tony Staw staw@marvin.enet.dec.com James VanBokkelen jbvb@ftp.com John Veizades veizades@apple.com Steve Willis swillis@wellfleet.com John M. Wobus jmwobus@suvm.acs.syr.edu David Paul Zimmerman dpz@convex.com 2