miox interface

"Dean D. Throop" <throop@dg-rtp.dg.com> Fri, 29 May 1992 19:31 UTC

Received: from nri.nri.reston.va.us by ietf.NRI.Reston.VA.US id aa01660; 29 May 92 15:31 EDT
Received: from nri.reston.va.us by NRI.Reston.VA.US id aa09983; 29 May 92 15:38 EDT
Received: from dg-rtp.rtp.dg.com by NRI.Reston.VA.US id aa09952; 29 May 92 15:37 EDT
Received: from walrus.rtp.dg.com by dg-rtp.dg.com (5.4/dg-rtp-proto) id AA11562; Fri, 29 May 1992 14:03:44 -0400
Received: by walrus (5.4.1/140.2) id AA26024; Fri, 29 May 1992 13:59:53 -0400
Date: Fri, 29 May 1992 13:59:53 -0400
From: "Dean D. Throop" <throop@dg-rtp.dg.com>
Message-Id: <9205291759.AA26024@walrus>
To: x25mib@dg-rtp
Subject: miox interface
Cc: kzm@hls.com

The currently draft X.25 and IP over X.25 MIBs specify that a 
machine with an X.25 interface will have two entries in the 
ifTable.  One for LAPB and one for the X.25 packet layer.  There will 
not be an entry for the IP interface to X.25.  

This doesn't match our implementation very well.  For us it would 
make more sense to have three interfaces, one for LAPB, one for 
X.25 packet layer, and one for the IP interface to X.25.  We have a 
problem when we have two X.25 subnetworks because the IP interface 
to X.25 will only have one IP address and yet may use both X.25 
subnetworks.  

Since our X.25 may carrying more that just IP traffic we have 
trouble determine what to do when setting ifAdmnStatus to down.  
Should this stop all IP traffic but still allow other X.25 traffic 
or does this stop all X.25 traffic?  For us it is a very different 
thing to stop X.25 than to stop the IP interface to X.25.  

We also use different names for the IP-over-X.25 interface and the 
X.25 interface.  Thus if ifDescr returned the name of the X.25 
interfaces, the system administrator wouldn't recognize them as 
IP-over-X.25 interfaces.  If we returned the name of the IP over 
X.25 interface, we wouldn't have a name for an X.25 interface in the 
case when we have one ip-over-X.25 connected to two X.25 
interfaces.  

Do you know of any reason why we can not have three interfaces?  

Would it make sense to explicitly say in the the MIBs that
implementations can allow either?  

Maybe an Ethernet-to-X.25-bridge would have interfaces for the 
Ethernet, LAPB, and X.25 while a Router should have interfaces for 
the Ethernet, LAPB, X.25, and IP-over-X.25.  What to the b-router 
vendors think?  Do b-routers need to have different interfaces to 
indicate when traffic is routed into X.25 then when traffic is 
bridged onto X.25?  

How do we use the ifType field to indicate this?  The current types 
ddn-x25(4) and rfc877-x25(5) really pertain to the IP interface to 
X.25 and have nothing to do with X.25.  We could say the X.25 
interface has a type of other(1) and let ifSpecific be the real 
type specifier.  Then the existing types become the type of the 
IP over X.25 interface.  

Comments?

Dean Throop			throop@dg-rtp.dg.com