velasco@beowulf.ucsd.edu (Gabriel Velasco) (04/11/90)
Quoting Andrew Tanenbaum:
The IEEE 802.3 standard is for a 1-persistent CSMA/CD LAN.
[stuff deleted]
The 802.3 standard has an interesting history. The real
beginning was Abramson's ALOHA system in Hawaii, to be
discussed later in this chapter. Later, carrier sensing
was added, and Xerox built a 2.94-Mbps CSMA/CD system to
connect over 100 personal workstations on a 1-km cable
(Metcalfe and Boggs, 1976; Schoch, 1987). This system
was called Ethernet after the luminiferous ether, through
which electromagnetic radiation was once thought to propagate.
[stuff deleted]
The Xerox Ethernet was so successful that Xerox, DEC, and Intel
drew up a standard for a 10-Mbps Ethernet. This standard
formed the basis for 802.3. The published 802.3 standard
differs from the Ethernet specification in that it
describes a whole family of 1-persistent CSMA-CD systems,
running at speeds from 1 to 10 Mbps on various media.
[stuff deleted]
Many people (incorrectly) use the name "Ethernet" in a
generic sense to refer to all CSMA/CD protocols, even though
it really refers to a specific product that implements
802.3.
Quoted from the second edition of _Computer Networks_ by Tanenbaum.
________________________________________________
<>___, / / | ... and he called out and said, "Gabriel, give |
/___/ __ / _ __ ' _ / | this man an understanding of the vision." |
/\__/\(_/\/__)\/ (_/_(/_/|_ |_______________________________________Dan_8:16_|grr@cbmvax.commodore.com (George Robbins) (04/11/90)
In article <8145@sdcsvax.UCSD.Edu> velasco@beowulf.UUCP (Gabriel Velasco) writes: > Quoting Andrew Tanenbaum: > > The IEEE 802.3 standard is for a 1-persistent CSMA/CD LAN. > [stuff deleted] ... > The Xerox Ethernet was so successful that Xerox, DEC, and Intel > drew up a standard for a 10-Mbps Ethernet. This standard > formed the basis for 802.3. The published 802.3 standard > differs from the Ethernet specification in that it > describes a whole family of 1-persistent CSMA-CD systems, > running at speeds from 1 to 10 Mbps on various media. Right, Andy does a pretty good job of telling the history. Now look at what you said before that I took exception to: >> The actual "Ethernet" hardware standard is called 802.3. Ethernet is a >> specific implementation of that standard that has become "standard" and is >> licensed out to various manufacturers. As I read this, you've inverted history. The 'actual "Ethernet" hardware standard' is a couple of references by DEC/Intel/XEROX. The corresponding reference for 802.3 compliant networks is an ANSI/ISO document. Ethernet is not a "specific implementation of that standard [802.3]", rather it is specifically a network compliant with the DEC/Intel/XEROX documents. Likewise "[Ethernet as a specific implementation of 802.3]" has not become a standard, rather it was an existing standard that was widely implemented *before* 802.3 existed. Strictly speaking DEC/Intel/XEROX "ethernets" are *not* 802.3 compliant because of a number of mostly minor technical issues and thus can't be viewed as a "specific implementation of 802.3" in the sense of being a proper subset. While the 802.3 standard credits the original DEC/Intel/Xerox standard and implementors, it does not make reference to that standard nor does it include it in some form of "grandfather clause". Now in a broader sense, Ethernet is a term that is used on a day to day basis that is used any of the 10 MB/s coaxial networks using baseband signalling and "ethernet" protocols. With further qualification it is used to describe other media (thin-wire, twisted-pair), broad-band (carrier based) or even networks consisting of "ethernet" segments connected via near-full-speed fiber optic or microwave links. It is not used to describe non-CSMA/CD links such as token-ring and generally not used to with reference to networks like "Star-LAN" which uses "ethernet" style low-level signalling/protocols, but over a different media at a different data rate (twisted-pair/~1MB/s). The 802.3 standard and it's addendum does address most of the alternate network forms of the first group and in some ways it is better than the original "ethernet" standards. The one feature that hasn't been widely adopted is the redefinition of the original "ethernet type field" as a "length field". This is not at all a bad idea, since it lets one predict when the packet ends, rather than relying on some indirect evidence that you've received the whole packet, however since the majority of the installed base doesn't share this interpretation it's not often used (at least in the traditional baseband "ethernet" context). The final issue is that when speaking outside of ones area of direct expertise, it is a good idea to prefix otherwise "absolute" statements with some kind conditional phrase the indicates your level of confidence or the source of your knowledge. In this way, the reader is alerted that if the matter is crucial then they should verify it themselves and also if someone feels the information is incorrect, it is a lot easier to let something slide that begins with "I believe" or "As I understand it" then an apparent statement of absolute fact. -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing: domain: grr@cbmvax.commodore.com Commodore, Engineering Department phone: 215-431-9349 (only by moonlite)