Talk:Ethernet/Archive 2

From Wikipedia, the free encyclopedia

Archive This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Contents

Official logos?

Are there any official logos for the various versions and types of ethernet, like there are for Firewire and USB? Or did nobody think of things like that in the 1970's? —Preceding unsigned comment added by Bizzybody (talk • contribs) 07:05, August 27, 2007 (UTC)

Not that I know of. One difference between Ethernet and the others you mentioned is that those others are semi-proprietary, owned and marketed by industry consortia. Ethernet was an open standard from day 1. As I understand it, the name was at one time a trademark of Xerox, but it was released into the public domain around the time the first Ethernet standard came out. (I don't have a reference for that; it would be interesting to try to track that down.) Paul Koning 13:53, 27 August 2007 (UTC)

Broken link?

The ACM classics page appears to no longer list the original ethernet paper. —Preceding unsigned comment added by 130.217.250.13 (talk) 01:21, 25 September 2007 (UTC)

"Dealing with multiple users" confusing/misleading

The first part of this section "CSMA/CD shared medium Ethernet" cofuses me. It begins "Ethernet originally used a shared coaxial cable... (CSMA/CD) governed the way the computers shared the channel..." and then carries on in the past tense (apart from the final paragraph oddly enough).

It implies that the CSMA/CD method is no longer used and then provides no information on the currently used method of dealing with multiple users :/

Is it just bad wording?

Thanks in advance.

Cre8tor 20:35, 28 October 2007 (UTC)

Coax is pretty much obsolete -- it only suppors the original 10 Mb/s version of Ethernet, and it is inconvenient to install. CSMA/CD carries over into the predominant twisted pair version of Ethernet, though. While switches (bridges) are now very common, Ethernet hubs still exist. If you have one of those, then you're using CSMA/CD. Paul Koning 13:49, 29 October 2007 (UTC)
Quoted from the article
"Despite the physical star topology, hubbed Ethernet networks still use half-duplex and CSMA/CD, with only minimal activity by the hub, primarily the Collision Enforcement signal, in dealing with packet collisions. Every packet is sent to every port on the hub, so bandwidth and security problems aren't addressed. The total throughput of the hub is limited to that of a single link and all links must operate at the same speed."
"When a twisted pair or fiber link segment is used and neither end is connected to a hub, full-duplex Ethernet becomes possible over that segment. In full duplex mode both devices can transmit and receive to/from each other at the same time, and there is no collision domain."
So if you keep reading it does indeed explain how multiple users are dealt with nowadays. Plugwash 20:34, 29 October 2007 (UTC)

TP hubs are little computers with multiple NICs

"When a twisted pair or fiber link segment is used and neither end is connected to a hub, full-duplex" is nonsense. A hub is like a PC with multiple NICs in crossover mode and set to forward packets. Half duplex is required on coax, not with separate receive and transmit twisted pairs. Some older NIC chips were too slow or not enough ram or bad drivers, ran half duplex. Again, the TX/RX wires don't touch, collisions are actually handled in Hub firmware.

Shjacks45 (talk) 17:27, 9 December 2007 (UTC)

Sorry, but you're mistaken. A hub is not a computer; it's an Ethernet repeater. An Ethernet with repeaters runs in half duplex mode, always. Also, half duplex has nothing to do with "too slow" NIC chips; it's one of the two operating modes of Ethernet (the original one). Please see the Ethernet standard for more details. Paul Koning (talk) 00:50, 10 December 2007 (UTC)

correct "related" 100BaseVG

Actually 100Base-VGAnylan was a Hewlett-Packard token passing protocol. Wiki on IEEE 802.12:"100BaseVG-AnyLAN technology developed by Hewlett Packard, which uses the demand priority access method." 'Overview of IEEE 802.12 Demand Priority'

Shjacks45 (talk) 17:59, 9 December 2007 (UTC)

Interesting. The article doesn't call it "token passing". It was originally proposed as a faster Ethernet (competing with 100BaseT, which came out the winner) so it certainly is related in that sense. Calling it "token passing" would have been a marketing mistake even if it's technically correct (which I'm not convinced about) -- it would make it sound like 802.4, which was even back then obviously a complete dead end. Paul Koning (talk) 00:55, 10 December 2007 (UTC)

looking for DLC??

win31/win95 to print to HP JetDirect Card network printers directly you had to install HP Jetadmin printing client and "DLC" protocol. ref MS KB Q117629 Q119068 Q94084 Q96623 but most 9x and w31 stuff deleted. If you've an old CD, the (Q146238)Admin.doc file included with Microsoft

Windows 95 Service Pack 1 (has IE 2.0): "Microsoft 32-Bit DLC Protocol

for Windows 95", about access to IBM SNA stuff. IBM incuded it with OS/2. Also listed by HP as DLC/LLC.

DLC is seminal in that it is a MAC address to MAC address protocol. Frustrating that documentation for old software disappears. However in setting SNA connectivity up on a reinstall, some NIC cards worked (like 3com) and some (ne2000 clones) didn't until Microsoft Netbeui was added. TID and MS docs said it was missing functionality in NIC firmware.

Shjacks45 (talk) 21:32, 9 December 2007 (UTC)

802.11

I removed the link to "Wireless Ethernet"(aka 802.11 aka WiFi) in the "Related Standards" section. It stated that 802.11 uses the Ethernet headers(which is false) and that 802.11 is interoperable with Ethernet(which is ridiculous -- how can a standard for wired communication with interoperable with a standard for wireless communication?)--Ryan Stone 18:13, 25 Nov 2004 (UTC)

Wrong, to the OS, 802.11 is ethernet with ethernet frames, turn on network bridge in WinXP to see, or you can explain how on earth this [1] and this [2] possible? —Preceding unsigned comment added by Patcat88 (talk • contribs)
There are a lot of networking technologies (like Ethernet, 802.11, Token ring, etc.), and a plethora of devices for converting from one to another, allowing them to interoperate. But without some kind of bridge to convert the media and frame types, devices using these different technologies can't communicate. And the bridging can get tricky; consider that 802.11 allows longer frames than Ethernet does, and that 802.11 must use IEEE 802.2. --Rick Sidwell 22:25, 12 Feb 2005 (UTC)
802.11 is not ethernet with ethernet frames to the OS. This is a feature of the NIC driver, or as with Windows and its miniport drivers, it's a function of the 802.11 MAC-layer driver in NDIS. The NDIS spec states that the the MAC-layer driver should present NDIS with 802.3 frames, thus it performs a conversion. If one has a driver with so-called monitor mode support, this functionality can be turned off. EOD :-) --Gosub 13:08, 24 Jun 2005 (UTC)
802.11 was designed to serve as a wireless extension to Ethernet rather than as a fully independent and competing standard, so it's not inaccurate to refer to it as Wireless Ethernet. There is so much historical crap in this article about Coax Ethernet that it's probably best not to make it any longer than it already is.
I mean, given that nobody has used a coax-based Ethernet for ten years, why not take all this collision detection crap and move it to another article? RichardBennett 02:54, 4 June 2006 (UTC)
I beg to differ. There were still buildings at MIT using thicknet as recently as three years ago (and may still be for all I know); there are probably many academic institutions where this was the case. But you are probably right that this discussion should probably move to a History of Ethernet article. 121a0012 03:02, 4 June 2006 (UTC)
I've personally never encountered thicknet. Thinnet on the other hand was arround much more recently and i know at least one person who still runs it today. In any case hubs (which are still sold new) still rely on that collision detection capability and i know of a recently released embedded ethernet controller that doesn't support autonegotiaton (limiting it to half duplex if the other end is an unmanged switch). Full duplex may be becoming more common but CSMA/CD will be with us for some time yet. In summary you can't really understand hubs without understanding the shared medium and to only describe switched ethernet would produce a hugely distorted picture even of current ethernet practice. Plugwash 08:10, 6 June 2006 (UTC)
"I mean, given that nobody has used a coax-based Ethernet for ten years, why not take all this collision detection crap and move it to another article?" collision detection is also important on hub based ethernet which is still pretty common on existing installations. Plugwash (talk) 21:55, 22 February 2008 (UTC)
I agree! Electron9 (talk) 22:19, 22 February 2008 (UTC)

Preamble and SFD was incorrect

The article originally stated that the preamble is 7 octets of 01010101 and the SFD is 11010101. The correct preamble is 7 octets of 10101010 and the SFD is 10101011. I have modified the article to reflect this. http://docs.hp.com/en/98194-90053/ch02s04.html confirms this. —Preceding unsigned comment added by Blutrot (talk • contribs) 05:58, 2 January 2008 (UTC)

Is that "10101010" and "10101011" in the sense that the preamble and SFD go onto the wire as 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1? If so, that means that the actual octets of the preamble are 0 1 0 1 0 1 0 1, at least as I interpret 3.3 "Order of bit transmission" in IEEE Std 802.3-2005, which says "Each octet of the MAC frame, with the exception of the FCS, is transmitted low-order bit first." Or is that "10101010" and "10101011" in the sense that those are the octets, so that the preamble and SFD go onto the wire as 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 1 0 1 0 1 0 1? Guy Harris (talk) 09:16, 2 January 2008 (UTC)
Yes. The source of the problem is that IEEE 802 writes octets with the LSB on the left. In other words, you have to read them as bitstrings, not as numeric values.
When dealing with bit patterns that aren't really numeric values -- like the preamble/SFD -- this is confusing but at least not completely nuts. It gets worse, though, when you try to make sense out of the LLC standard (802.2). It deals very clearly with octets, and in fact those are transmitted in whatever order the specific MAC used transmits them. But still they are written LSB first.
Paul Koning (talk) 12:04, 2 January 2008 (UTC)
I will review some more documentation and attempt to make sense of what actually goes on. On nodes (not switches) running 10BASE-T and 100BASE-TX, pins 1 and 2 are used to transmit while pins 3 and 6 are used to receive. Perhaps it would make sense to say pin 1 transmits X and pin 2 transmits Y (x and y being whatever they transmit) 31 times, followed by pin 1 and 2 both transmitting 1 once. Although, this isn't as helpful with 1000BASE-T which uses four wires to transmit and receive.
Blutrot 18:33 2 January 2008 (UTC)
I tested with Preamble=0101 and SFD=1101 (as 4'b1101 in verilog):
       0x0000:  ffff ffff ffff 1234 5678 9abc 0000 1083  .......4Vx......
       0x0010:  4567 6767 6767 6767 6767 6767 6767 6767  Eggggggggggggggg
       0x0020:  6767 6767 6767 6767 6767 6767 6767 6767  gggggggggggggggg
       0x0030:  6767 6767 6767 6767 6767 6767 c948 5581  gggggggggggg.HU.
And with Preamble=1010 and SFD=1011
       0x0000:  4567 6767 6767 6767 6767 6767 6767 6767  Eggggggggggggggg
       0x0010:  6767 6767 6767 6767 6767 6767 6767 6767  gggggggggggggggg
       0x0020:  6767 6767 6767 6767 6767 6767 467a 1c70  ggggggggggggFz.p
Both packets are sent out identical with the exception for an incremental 32 bit counter value, and the 32 bit crc. I have also verified that 4'b1000 sent to PHY will set TXD3=1 and the rest =0 etc.. I think the conclusion must be that the use of 0101 and 1101 with MSB at leftmost position is the correct ones. The only reason for the HP values must be they are using LSB at leftmost position.
As for TX pins, they are differential, so pin 1 transmitts X, then pin 2 transmitts the opposite of X to produce the differential signal. X usually being a voltage in the 3.3V range. 1000Base-T uses signal cancellation schemes to make simultaneous tx & rx work. Electron9 (talk) 19:08, 2 January 2008 (UTC)
Ok, so I got confused. The IEEE 802.3 spec is the authority here. (FWIW, the HP doc has the same pattern but without the explanation to make it intellegible.) The preamble is 101010... with the leftmost bit sent first. (It even adds "It should be noted that the preamble ends with a '0'.") And the SFD is 10101011 again with the leftmost bit sent first. So you have to read them as bitstrings, not as binary integers.
If you want to write everything as octets (binary 8-bit integers) then you'd reverse the strings because LSB is sent first in Ethernet -- and that brings you to the text in the article as it stands right now.
Paul Koning (talk) 20:21, 2 January 2008 (UTC)
Interestingly enough the description of what the SFD looks like for 100 Mb Ethernet was wrong, since the nibble order is low to high. Fixed. 20:25, 2 January 2008 (UTC) —Preceding unsigned comment added by Paul Koning (talk • contribs)
One could argue that Ethernet is little endian :-)... (hello Intel?). Anyway, any link to that IEEE 802.3u document that you reference too?, IEEE isn't letting people have it easily. Electron9 (talk) 21:21, 2 January 2008 (UTC)
I see what you are saying. I apologize for the confusion this caused in the article. Thank you for all of your insightful comments.
Blutrot 23:15 2 January 2008 (UTC)
802.3u was the Fast Ethernet spec, according to the IEEE 802.3 page. The IEEE does, in fact, make it harder to get old standards than current ones - the older standards aren't directly available from the Get IEEE 802 page, but that page points you to ShopIEEE for "New, Superseded, and Withdrawn standards".
802.3u's contents are probably mostly if not entirely found in the current 802.3 standard; see 802.3-2005 section 2 (the standard is in 5 separate PDFs on the Get IEEE 802 site). Guy Harris (talk) 23:36, 2 January 2008 (UTC)
Exactly. In fact, all of the 802.3 standards are there. What's been retired is the individual titles as separate documents, not the substance. And they are free -- so they the ideal sources to cite in this article... Paul Koning (talk) 11:58, 3 January 2008 (UTC)


looks like 802.3 IEEE standard is available for download free IEEE 802.3 part 1 page 50.
"The SFD field is the sequence 10101011. It immediately follows the preamble pattern and indicates the start of a frame."
also page 105 states
"The preamble pattern is: 10101010 10101010 10101010 10101010 10101010 10101010 10101010 The bits are transmitted in order, from left to right"
My suggestion is to follow the standard - and display the byte pattern Kevmcs (talk) 23:21, 27 January 2008 (UTC)
The problem with just following the standard is that the standard is seriously and unnecessarily confusing. I restored the "byte wise" way of showing the pattern, but added a paragraph explaining that the IEEE standard has it the other way around. One reason for showing it byte wise is that it's the only way to make sense out of what happens in the 100 Mb/s (4b/5b code) or faster (8b/10b code) transmission. Paul Koning (talk) 23:37, 27 January 2008 (UTC)
The bit pattern end with 11 is important to display as last 2 bits of the eight byte are always consecutive 11. Even if the preamble had been corrupted, the following destination address can be identified correctly. The table needs to reflect that the last two bits are 11. The authoritative source is 802.3 IEEE standard, and we need to source it and table should reflect it Kevmcs (talk) 23:56, 27 January 2008 (UTC)

Introduction year & Name in all caps..?

Paul!

  • ) Should the pre 3 Mbps LAN networks count as "ethernet" which were introduced in 1980, or should one define it as a minimum 10 Mbps LAN that were introduced first in 1985 ..?
  • ) Is there any official standard that backs up the claim that the standard names, like 10BASE5 etc. Should be in all uppercase ..?, anyway I find it much more readable in 'Base' format rather than 'BASE'.

Electron9 (talk) 17:53, 22 February 2008 (UTC)

In answer to the first question, given that the paper "Ethernet: Distributed Packet Switching for Local Computer Networks" describes a 3 Mbps LAN, I'd say that the term "Ethernet" covers such a LAN. However, that LAN wasn't compatible with DIX Ethernet (the source and destination addresses were 8 bits and LAN-specific, rather than 48 bits and universal), so, while it is a form of Ethernet, it should be noted as being a different form of Ethernet from the DIX Ethernet and 802.3 Ethernets that succeeded it.
In answer to the second question, yes, there is an official standard which calls them "10BASE5", "10BASE2", "10BASE-T", "100BASE-T", etc. - it's called IEEE 802.3-2005. Guy Harris (talk) 18:37, 22 February 2008 (UTC)
About year of introduction -- the Ethernet V2 spec is dated 1982. I believe that's the document that established the Ethernet standard and started the stream of products. I'm not sure what's in the V1 document or how it differs. As for the 3 Mb/s Ethernet, that certainly was called Ethernet, but it wasn't a product, only a research prototype at Xerox PARC. So when speaking of Ethernet as a commercial technology, my inclination is to use 1982 as the start year. To make the picture complete, it would make sense to discuss the research stage work, which includes the 3 Mb/s version. Paul Koning (talk) 15:56, 25 February 2008 (UTC)
Split the article in 'Ethernet I' and 'Ethernet II' then ..? (or how to deal with this?) Electron9 (talk) 16:58, 25 February 2008 (UTC)
No, just leave the article as it is, with the Xerox experimental Ethernet mentioned in the History section. That is part of the history of Ethernet. Guy Harris (talk) 18:18, 25 February 2008 (UTC)

Duplication

Ethernet#Autonegotiation_and_duplex_mismatch and Ethernet_over_twisted_pair#Autonegotiation_and_duplex_mismatch contain a lot of duplicated text and should be merged. JonathanWakely (talk) 14:34, 6 March 2008 (UTC)