Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!apple!csl!hercules.csl.sri.com!forster
From: forster@dustbin.cisco.com (Jim Forster)
Newsgroups: comp.sys.novell
Subject: Re: Encapulation of Novell IPX packets over TPC/IP networks?
Message-ID: <FORSTER.91Jan6230825@dustbin.cisco.com>
Date: 7 Jan 91 07:08:25 GMT
References: <1044@rsp.UUCP> <11090002@acf3.NYU.EDU> <19477@netcom.UUCP> <2546@excelan.COM>
Sender: news@csl.sri.com (NEWS Administrator)
Reply-To: forster@cisco.com (Jim Forster)
Distribution: comp
Organization: cisco Systems, Inc.
Lines: 48
In-Reply-To: brian@ca.excelan.com's message of 29 Dec 90 00:24:46 GMT

In article <2546@excelan.COM> brian@ca.excelan.com (Brian Meek) writes:

   Configuring the IPX over IP router involves telling the LAN driver how
   to handle IPX broadcast requests by supplying a list of IP addresses
   where peer IPX routers exist.  When IPX hands a broadcast packet to the
   SK-Net IPX/IP LAN driver, the driver wraps it up in a UDP datagram and
   copies it to each IP address in the peer list.  One can define a list of
   up to 20 IPX/IP peers, last I checked.  Of course, all servers / or
   external _IPX_routers_ must agree as to the IPX network address of the
   "Tunnel".

This is one of the scaling problems -- manual configuration.  Not too bad
for a handful of tunneling routers, but even if the limit of 20 were boosted,
its not a good solution for a large network.

   >The advantage? Now you can use "real" routers like Cisco or Wellfleet, add
   >HDLC, SLIP and PPP links, router management via SNMP etc, and best of 
   >all - get rid of RIP.
   >
   Get rid of RIP?  I assume you're referring to IPX-RIP since tunneling IPX 
   within UDP has nothing to do with the IP routing protocol in use.  Anyway,
   IPX RIP doesn't go away with this encapsulation scheme, but it does give
   one the ability to eliminate the broadcasting of RIP packets across IPX
   internets by hiding all IPX broadcasts in UDP unicasts targeted at specific 
   IP addresses belonging to NetWare servers or IPX routers (good luck parsing
   that one :-).

I think Brian's got it, but to be more specific about the problems, run the
numbers: if there are 20 tunneling routers, then every 30 seconds each
sends a RIP update to all the others, by sending 19 separate packets
accross the IP net.  And it also receives 19 packets.  So the 1 RIP packet
per 30 seconds turned into about 20 every 30 seconds, or almost 1 per
second.  And the IP net sees about 12 per second (20*20/30).

Stated another way, this is an N**2 (N-squared) solution, which is not good.

   BTW:  Novell views this as a reasonable approach for utilizing existing
   SPX/IPX based applications and connecting NetWare workgroups together 
   across IP internets. 

Well, I suppose if thats all you can manage, it certainly would be useful,
but with "'real' routers" and native IPX packets, its much more
straightforward, and you still have SNMP, HDLC & FDDI links, etc.



Jim Forster
cisco Systems