IMHO Both proposed options can reuse existing standards and code. Both
should work.
What's there to gain by making the nonce shorter?
What's there to gain by limiting the nonce to a fixed length?
I'm just thinking of cases like a whole bunch of products with IPv6 RFID
tags going through a scanner on a palette and all trying to get an
address at once.
regards,
RayH
Message: 1
Date: Mon, 17 Oct 2011 17:41:05 +0200
From: Philip Homburg<[email protected]>
To: "Hemant Singh (shemant)"<[email protected]>
Cc: IPv6 WG Mailing List<[email protected]>
Subject: Re: FW: New Version Notification for
draft-hsingh-6man-enhanced-dad-01.txt
Message-ID:<[email protected]>
In your letter dated Mon, 17 Oct 2011 10:25:19 -0500 you wrote:
>Appreciate the quick reply. Note BrianC already noted that 20 bits will
>not suffice by saying "It puts you into birthday-paradox territory on a
>LAN with a few hundred nodes.". His email is at the URL below.
>
>http://www.ietf.org/mail-archive/web/ipv6/current/msg14835.html
I'm not sure how you would get a birthday paradox here. Applying the birthday
paradox, you would get that if you have a network with a 1000 nodes, and all
nodes try to configure the exact same IPv6 address at the same time, then
there is a probability of about 0.5 that two nodes will mistakenly assume that
there is a loopback situation and all the others will detect a duplicate
address.
That's not a problem I would worry about.
--------------------------------------------------------------------
IETF IPv6 working group mailing list
[email protected]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------
