metric 15 == infinity??

Jeffrey C Honig <> Tue, 05 July 1994 22:18 UTC

Received: from by IETF.CNRI.Reston.VA.US id aa10521; 5 Jul 94 18:18 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aa10517; 5 Jul 94 18:18 EDT
Received: from by CNRI.Reston.VA.US id aa04698; 5 Jul 94 18:18 EDT
Received: by id AA18362 (5.65c/UK-2.1-940401); Tue, 5 Jul 1994 18:16:57 -0400
Received: from MITCHELL.CIT.CORNELL.EDU by with SMTP id AA28047 (5.65c/UK-2.1-940401); Tue, 5 Jul 1994 18:16:47 -0400
Received: from (MITCHELL.CIT.CORNELL.EDU []) by (8.6.4/8.6.4) with ESMTP id SAA04571 for <>; Tue, 5 Jul 1994 18:13:39 -0400
Message-Id: <>
Subject: metric 15 == infinity??
Organization: Information Technologies/Network Resources; Cornell University, Ithaca, NY
X-Mailier: MH-E [version 4.1+] MH [version 6.8.1]
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Tue, 05 Jul 1994 18:13:38 -0400
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: Jeffrey C Honig <>

The way rfc1058.txt is written, when you receive a RIP route, you add
the interface cost to the metric in the packet to come up with the
metric to install in your routing table.  If this calculated metric is
16 or greater, you consider the route to be unreachable.  The cost is
1-15 with 1 being the default.  In pseudo code this is:

Now consider a received metric of 15.  With the default interface
cost, that results in a calculated metric of 16.  So the route will be
ignored.  Basically limiting valid RIP metrics to the range of 1-14
instead of 1-15.  I think this is broken.

I can see two potential solutions:

I) If the calculated metric is *greater* than 16, consider it deleted.
   If it is 16 or less, add it to the routing table.  Of course, do
   not send a metric of 16.

II) Use the packet metric to decide if the route should be added to
   the routing table.  Then use the calculated metric to compare
   against other routes.  So a route with a metric of 15 or less would
   always be considered for installation in the routing table,
   regardless of the interface metric.  This would result in routing
   table metrics in the range of 2-30.

I'd be interested in hearing other people's opinions.  And learn what
you've done in your code.