Re: DAD question

[Bob Hinden <[e-mail]>]

Steinar,

On Aug 11, 2012, at 9:01 AM, [e-mail] wrote:

>>> it may be a fair assumption that if an address based on the MAC address is duplicate, the MAC address itself is a duplicate.
>> 
>> True, but the odds of this happening are very low.  I wonder if we have any data on DAD detecting duplicate addresses and their cause.
> 
> You may need to qualify "very low".
> 
>> For example, has any seen any actual duplicate MAC addresses?  It would be good to collect some data.
> 
> Duplicate MAC addresses are regularly seen in the wild. As an example,
> from a nearby DHCP server, I have the following duplicates from a total
> of 65499 MAC addresses:
> 
> 5 00:90:4c:91:00:01;
> 3 00:40:10:20:00:02;
> 3 00:00:00:00:00:00;
> 2 bc:b1:f3:61:28:e3;
> 2 98:0c:82:84:ef:83;
> 2 34:21:09:03:76:f9;
> 2 00:1f:1f:8c:d4:d5;
> 2 00:1f:1f:8c:d4:cd;
> 2 00:1d:73:11:11:13;
> 2 00:11:22:33:44:56;
> 
> So from this sample a total of 10 MAC addresses occur more than once
> (different customers, different locations).

Interesting.  Are these different machines/interfaces with the same Mac addresses, or the same machine connecting via some other path or location.

My question was about machines/interfaces.  That is, are we seeing manufacturing mistakes, etc.?

Thanks,
Bob



> 
> I make no claims about these numbers being representative. My main 
> point is that duplicates *occur*, for several different reasons, e.g.
> 
> - Manufacturing mistakes
> - Software bugs
> - MAC address explicitly set
> 
> Steinar Haug, Nethelp consulting, [e-mail]