Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cwjcc!tut.cis.ohio-state.edu!ucbvax!husc6!bu-cs!kwe
From: kwe@bu-cs.BU.EDU (kwe@bu-it.bu.edu (Kent W. England))
Newsgroups: comp.protocols.tcp-ip
Subject: Re: Kenmore MicroWave Oven For Sale
Message-ID: <34957@bu-cs.BU.EDU>
Date: 19 Jul 89 22:21:43 GMT
References: <325@arnor.UUCP> <42958@bbn.COM>
Reply-To: kwe@buit13.bu.edu (Kent England)
Followup-To: comp.protocols.tcp-ip
Organization: Boston U. Information Technology
Lines: 43

In article <42958@bbn.COM> djw@BBN.COM (David Waitzman) writes:
>The microwave obviously speaks TCP:
>Here is the "traceroute" output to kenmore.kitchen.net from here:
>
>traceroute to kenmore.kitchen.net (224.1.54.1), 30 hops max, 40 bite packets
> 1  circe.bbn.com (128.89.0.247)  0 ms  0 ms  10 ms
> 2  128.89.0.1 (128.89.0.1)  60 ms  70 ms  70 ms
> 3  * * *
> 4  pc.livingroom.house.ny.us (225.9.5.4)  90 ms  90 ms  90 ms
> 5  fridge.kitchen.net (224.4.5.3) 140 ms  150 ms  140 ms
> 6  pc.livingroom.house.ny.us (225.9.5.4)  160 ms  170 ms  160 ms
> 7  fridge.kitchen.net (224.4.5.3) 300 ms  310 ms  300 ms
> 8  pc.livingroom.house.ny.us (225.9.5.4)  320 ms  330 ms  320 ms
> 9  fridge.kitchen.net (224.4.5.3)  340 ms  350 ms  340 ms
>10  kenmore.kitchen.net (224.1.54.1)  360 ms  370 ms  360 ms
>
>Note the temporary routing loop from the pc to the fridge.  Anyone have
>any ideas on why?
>
>-david


	I ran an snmpsrc on kitchen.net and found this:

bu-netcop{kwe}44> snmpsrc livingroom-gw kitchen-net
Copyright (C) 1988 NYSERNet INC

Looking for source of routing of kitchen-net....

gateway=livingroom-gw   next-hop=kitchen-gw     metric=3      remote rip
gateway=kitchen-gw      next-hop=bedroom-gw     metric=205871 remote criscoIgrp
gateway=bedroom-gw      next-hop=livingroom-gw  metric=277889 remote boudoir
receive timed out.  Retrying....
bu-netcop{kwe}45>

	Obviously this kind of appliance networking has some bugs to
be ironed out.  I understand that some appliances do not yet implement
split horizon, in particular some refrigerators.  Perhaps if they
tried a link state algorithm...  Fortunately they decided to do SNMP
(after fierce lobbying from Marty Schoffstall) and didn't wait for
CMIP or else we wouldn't have been able to figure things out this far.

	--Kent England, Boston University