[sllug-members]: mac address

[Lamont R. Peterson]

On Saturday 15 July 2006 21:45, Shaun Kruger wrote:
> > > (original ethernet used this 2 bytes as a
> > > packet length identifier).
> >
> > That is wrong.
>
> I may have just propagated hearsay that I had come to understand from
> reading about the first experimental ethernet (1972).  I had gathered
> that there wasn't a protocol ID field in 1972 ethernet as the Xerox
> protocol was the only one going over the wire.  The protocol ID was
> needed when it was being standardized so multiple protocols could
> share it.  I'm trying to figure out where I read this.  I may have
> also misunderstood something elsewhere.

I haven't found a reference about that (admitidly, I didn't look all that 
hard, either).

> This Wikipedia article is short, but it covers the use of the 2 bytes
> after src and dest mac addresses within a frame.
> http://en.wikipedia.org/wiki/DIX

It does talk about the 2 bytes following the MAC addresses, but it leaves out 
the rest of the story, which, IMHO, makes it a little misleading.  
Additionally, it says nothing of use about timeframes or relationships of the 
various standards we're discussing.  Sorry, but I don't think it's a good 
reference.

> It's nice being told you're wrong sometimes.

My apologies if I have offended you with the way that I phrased that.  At the 
time I was writing my response, I felt it important to be sure that I was 
completely unambiguous about the fact that your post was in error, so that 
others would not be misled by it's misinformation, either at this time or in 
the future (i.e. from the list archives).  It was never my intention to imply 
that you were in any way inferior to this topic.

> It makes you check your 
> sources.

The Wikipedia reference to which you provided a link is interesting.  Thank 
you for sharing it.

However, given the way that you snipped everything I wrote following my 
possibly rudely phrased line (again, I apologize if I offended) quoted above, 
the way you have phrased the response to it and the seeming implication that 
Wikipedia trumps the official RFCs which I quoted and referenced (following 
your snip point, I'm afraid), it would seem that you did not bother to read 
what I wrote.  I honestly can not tell for sure either way.  But if you did 
miss it (which would have been easy as, sometimes, my posts can ramble a bit) 
then you would have noticed that I described with a reasonable level of 
detail (I wouldn't say, "a great deal" there) the way in which both Ethernet 
II encapsulation and 802.3/802.2 encapsulation work, their relative positions 
in time and their relationship to each other on devices that support both 
frame formats.

In order to help with the timeframe (no pun intended :) ), the first IEEE 802 
standard was first drafted (or was it released, I never can seem to remember 
for sure) in February of 1980.  Interesting little tidbit:  I have often read 
or heard that the identifier "802" was derived by the fact that the meeting 
was in 1980, February (or month 2) and concatenated together.  However, 
according to the IEEE's own documentation, 802 was simply the next project 
number when the LMSC (LAN/MAN Standards Committee) held their first meeting.  
The LMSC is better known as IEEE Project 802 and, in fact, the name LMSC 
didn't come into being until some years after 1980 (these facts come from 
documents available on IEEE's website).

Incidentally, the IEEE Project 802 Standards documents are becoming available 
for free download.  Just visit [ http://standards.ieee.org/getieee802/ ] for 
the documents.  You can find the official IEEE Project 802 website at either 
[ http://grouper.ieee.org/groups/802/index.html ] or 
[ http://www.ieee802.org/ ] (I did run into a few broken links; mostly, it 
looked like "index.htm" needs to be "index.html" kind of errors).

In case anyone who is reading this does not feel like reading through the RFCs 
at this moment (not that I would blame you for choosing to stay awake ;) ), 
here are their ratification (as RFCs, that is) dates:

"RFC 894 - Standard for the transmission of IP datagrams over Ethernet 
networks" April 1984
"RFC 1042 - Standard for the transmission of IP datagrams over IEEE 802 
networks" February, 1988

You'll note from the titles that these RFCs apply to IP over Ethernet.  In 
other words, Ethernet II and 802.3/802.2 were around before these dates.  
However, in the context of the OP's question, these were the correct 
references.  Although Xerox PARC began formulating and using (experimentally) 
the predecessor of Ethernet in the early 1970's, it was not codified until 
1980 (with input from DEC and then Intel).  Ethernet II was released in 1982.  
The IEEE 802.3 standard was not completed until 1985 (which only includes 
Ethernet over thick-net and thin-net coax (I made such cables, back in the 
day).  It wasn't until the 1990 update to 802.3 that the standard supported 
10Base-T with UTP cabling.

The third RFC which I referenced earlier was "RFC 1122 - Requirements for 
Internet Hosts - Communication Layers" from October of 1989.  In this RFC, 
the integration of both Ethernet II and 802.3/802.2 for IP stacks which wish 
to provide support for both (among other things) is described.  It does this 
by building on top of both RFC 894 & RFC 1042 (among others).  Also, it is 
one of three RFCs which, "...defines and discusses the requirements for host 
system implementations of the Internet protocol suite."

BTW, according to the 802.2 LLC (Logical Link Control) sub-layer specification 
(which covers some stuff related to the OSI Layer 1 bits of Ethernet), the 
Ethernet frame is preceeded by 62 bits (or 7 bytes, I'm not sure which) of 
preamble (alternating 1's & 0's so the two NICs can achieve correct 
transmit-receive synchronization) and a 1 byte field called SFD (Start Frame 
Delimiter) whose value indicates whether this is going to be an Ethernet II 
frame or an 802.3 frame or an 802.3/802.2 frame or ... whatever.  
Interesting, not really related to answering the questions at hand and it's 
about Layer 1, which doesn't really affect the OP's question.

Thank you, Shaun, for getting me to "...check [my] sources," again.  But just 
in case anyone was wondering, I did check the RFCs the first time to make 
sure I was reffering to the right ones.  This time, I read through most of 
them, skimming sections that didn't look related to these questions or that 
were going into deeper detail that I didn't need right now.  I've also 
started reading the full IEEE 802 set of standards that are available for 
free download.  It's very interesting stuff, even if some of it reads like a 
cure for insomnia.

Perhaps I should go read RFC 1149 again 
[ http://www.faqs.org/rfcs/rfc1149.html ], that's a good one and it's short, 
too :) .  If you've never read RFC 1149 before, I would recommend it.  Just 
make sure you don't look at the date until you've finished reading it first.
-- 
Lamont R. Peterson <lamont at gurulabs.com>
Senior Instructor
Guru Labs, L.C. [ http://www.GuruLabs.com/ ]
GPG Key fingerprint: F98C E31A 5C4C 834A BCAB  8CB3 F980 6C97 DC0D D409
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://sllug.org/pipermail/sllug-members/attachments/20060716/d9da7315/attachment.pgp