[AusNOG] Internal use MAC addresses

[Mark Smith]

>________________________________
> From: Wade Roberts <ausnog at acquired-taste.net>
>To: Keith Sinclair <kcsinclair at gmail.com> 
>Cc: "ausnog at lists.ausnog.net" <ausnog at lists.ausnog.net> 
>Sent: Saturday, 23 February 2013 10:08 PM
>Subject: Re: [AusNOG] Internal use MAC addresses
> 
>
>Not quite, setting the 'next-to-least-significant-bit of first octet of OUI', or the 7th bit, to 1 would make 02-00-00 the simplest example of this rule.
>

Another way to look at it is that in hexadecimal, 0x02 needs to be present in the first octet. If you have a hexadecimal to binary calculator, convert the first octet from hexadecimal to binary and then look for a 1 in the Y position - xxxx xxYx.

For those interested in the origins of why MAC addresses are 48 bits and why they were intended to be globally unique, have a read of the following paper:

"48-bit Absolute Internet and Ethernet Host Numbers"
http://ethernethistory.typepad.com/papers/HostNumbers.pdf


Interestingly, they were originally intended to be globally unique host rather than interface identifiers, and if a host had multiple interfaces, the same MAC address was intended to be used on all of them. That would have precluded the attachment of two of the host's interfaces to the same link. I think the convention of a MAC address being assigned to an interface probably came about because when a host didn't already have a MAC address, it was easier for a network card manufacturer to put one on some ROM/EEPROM on the card, and then for manufacturing simplicity, do it on all cards rather than making it optional.

For example, the following shows that SUN originally adopted the same MAC address on all interfaces model and made it the default:

http://docs.oracle.com/cd/E19963-01/html/821-1458/geyqe.html#eyprp#scrolltoc