Adj Index Wed, 26 August 1992 12:59 UTC

Received: from by IETF.NRI.Reston.VA.US id aa01469; 26 Aug 92 8:59 EDT
Received: from NRI.NRI.Reston.Va.US by IETF.NRI.Reston.VA.US id aa01465; 26 Aug 92 8:59 EDT
Received: from by NRI.Reston.VA.US id aa05954; 26 Aug 92 9:01 EDT
Received: by; id AA29422; Wed, 26 Aug 92 06:00:33 -0700
Received: by; id AA13185; Wed, 26 Aug 92 05:41:53 -0700
Received: by; id AA13181; Wed, 26 Aug 92 05:41:52 -0700
Received: by; id AA28402; Wed, 26 Aug 92 05:41:51 -0700
Received: by (5.57/ULTRIX-fma-071891); id AA05085; Wed, 26 Aug 92 08:44:26 -0400
Message-Id: <>
Subject: Adj Index
Date: Wed, 26 Aug 92 08:44:25 -0400
X-Mts: smtp


Thanks for your note, there has been a lot of mail on this topic over the past
week, but the conclusion is (and I restate it here for final sanity check
since we have been around this issue so much):

	1.  When we turn the document to draft (there is 1 typo I know needs
to be fixed) we will also include the old table with the status of 'obsolete'

	2.  We will introduce a new table which will be identical to the one
you implemented except for the changes necessary for the index.  Specifically
it will be indexed as you suggest:

	{ phivAdjCircuitIndex,phivAdjNodeAddr }.

------- Forwarded Message

Received: by (5.57/ULTRIX-fma-071891);
	id AA05046; Wed, 26 Aug 92 08:09:42 -0400
Received: by; id AA26435; Wed, 26 Aug 92 05:06:52 -0700
Received: from mips.uucp by via EUnet with UUCP (5.64+/8+bit/IDA-1.2.8)id AA16609; Wed, 26 Aug 92 14:06:35 +0200
Received: from sony by (5.61/SMI-4.1)id AA09605; Wed, 26 Aug 92 15:04:01 +0200
Received: by (4.0/SMI-4.1)id AA12785; Wed, 26 Aug 92 14:03:46 +0100
From: (Sven Olav Lund)
Message-Id: <>
Subject: Adjacency index
To: (DECnetIV MIB Working Group)
Date: Wed, 26 Aug 92 14:03:45 WET DST
X-Mailer: ELM [version 2.3 PL11]
X-Charset: ASCII
X-Char-Esc: 29


I discussed the use of index for the Adjacency Group with Jon in
May 92. My suggestion was the double-index architecture which seems
to be the result of your discussion: 

	{ phivAdjCircuitIndex,phivAdjNodeAddr }.

Jon's reply lead me to implement the table with the single integer
index. The index is designed similar to the index proposed by Art:

	((CircuitIndex * 65536) + AdjNodeAddr).

We have already implemented the DECnetIV-MIB (RFC1289) in our router
products, which we are releasing very soon. I still agree that the 
double-index architecture is the best solution, so I will modify
our implementation to conform with the drafted MIB specification.

Sven Olav Lund, Dowty Network Systems.

------- End of Forwarded Message