Talk:Token ring

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Networking / Hardware (Rated Start-class, Low-importance)
WikiProject icon This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Low  This article has been rated as Low-importance on the project's importance scale.
Taskforce icon
This article is supported by Networking task force (marked as Mid-importance).
Taskforce icon
This article is supported by Computer hardware task force.
 
WikiProject Telecommunications (Rated Start-class)
WikiProject icon This article is within the scope of WikiProject Telecommunications, a collaborative effort to improve the coverage of Telecommunications on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the quality scale.
 ???  This article has not yet received a rating on the importance scale.
 

Proposed merge[edit]

I would like to suggest that this article merge with IBM Token ring entry. While much of this material is present in the other larger article - welcome your reference sources and further additions tomake a stronger Wikipedia entry. User:Beatgr 16:10, 24 November 2006 (UTC)

  • Support this merge. I think the uncapitalised name (ie this one) is the better article title. Andrewa 23:21, 8 December 2006 (UTC)
    • I would like to see the name usage per the IEEE standard (802.5), unless this newly merged article will be of broader context -- covering both the evental IBM implementations, IEEE 802.5 standard development and earlier versions developed before both (Proteon, Apollo, etc.) Beatgr 4:30, 11 December 2006 (UTC).

Correct name[edit]

See also Talk:IBM Token ring#Token-Ring - correct naming?. Andrewa 23:38, 8 December 2006 (UTC)

It seems that only IBM calls it "Token-Ring"; everywhere else, including the IEEE 802.5 spec, it's referred to as "Token Ring" or "token ring". So I think it's only appropriate to use in contexts where one speaks about IBM's implementation and not the protocol. I will create redirects for the different capitalizations once the requested move has been decided upon. -- intgr 17:44, 9 December 2006 (UTC)
Sounds good to me. Andrewa 19:00, 9 December 2006 (UTC)
The following discussion is an archived debate of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section.

The result of the debate was PAGE MOVED per discussion below. -GTBacchus(talk) 05:44, 15 December 2006 (UTC)

Requested move[edit]

IBM token ringToken ring — It is hardly ever called "IBM token ring", and whether the standard was originally designed by IBM or not, the article's title should remain neutral, as it talks about the IEEE 802.5 standard and not IBM's implementations. intgr 16:50, 9 December 2006 (UTC)

Survey[edit]

Add  * '''Support'''  or  * '''Oppose'''  on a new line followed by a brief explanation, then sign your opinion using ~~~~.
  • Support. As it now stands and is IMO likely to remain, the article is indeed about token ring architecture in general, not just the original IBM version of it. Andrewa 18:59, 9 December 2006 (UTC)
  • Support. more common term. --Akhilleus (talk) 06:36, 10 December 2006 (UTC)
  • Support - I've never heard it with IBM in it before. -Patstuarttalk|edits 14:39, 12 December 2006 (UTC)

Discussion[edit]

Add any additional comments:
Oh well, I put a lot of time into merging the IBM Token ring article into this one, then eliminating a ton of double and (triple!) redirects all through the encyclopedia. If people want to move the article back to simply Token ring, it's fine with me. I would ask that the history of everybody's contributions be kept, though, as I did when I merged IBM Token ring here. As for straightening out all the redirects, I'm going to stop now until the final decision on what to call this article is made. At least I got to tour a lot of the encyclopedia I had never seen before (rueful smile). Casey Abell 17:14, 9 December 2006 (UTC)
Your changes will not be lost, and the change history is always kept across moves. A page move is just a rename — the title will change, the content does not. (Or did I misunderstand your concern?) -- intgr 17:32, 9 December 2006 (UTC)
Oh, I can see that you have been rewriting ("piping") links in lots of different articles, and I'm sorry that your effort could go to waste. While changing redirects to redirects (that is, double-redirects), and links to disambiguation pages is very much appreciated, I don't think it's very productive to pipe/rewrite links just to avoid a single redirect – unless the link text (capitalization, etc) itself is incorrect and is changed, not piped, in the process. Link pipes also make very hard to track down the latter. -- intgr 17:56, 9 December 2006 (UTC)
Agree. Wikipedia guidelines should IMO be far more restrictive on the use made of pipes - not that people pay all that much attention to the restrictions already there, for example I often see pipes on disambiguation pages, which is explicitly frowned upon. Andrewa 19:07, 9 December 2006 (UTC)
Really good job on the merge, just BTW. Hang in there! Andrewa 19:10, 9 December 2006 (UTC)
My Wiki-conscience got the best of me. After I did the merge (appreciate the thanks!) I looked at the "What links here" tab and saw an unholy mess of double and triple redirects. I should have just left them alone, but I nagged myself about the policy on fixing redirects after a merge or move. So I started to straighten them all out, which took me to corners of this encyclopedia I never knew existed (wink). At least I learned more about computer networking than I ever wanted to know. (By the way, I may be one of the few WP editors old enough to have actually worked on a LAN that used token ring.) But whatever we end up calling this article, we probably should straighten out the mess of links to it. I know, it's thankless and dull work, and casual users may never be aware of it. But it keeps the bookkeeping a little neater. Casey Abell 20:11, 9 December 2006 (UTC)
You seem to have misunderstood us.
  • Sorting out double-redirects (that is, editing redirect pages that link to a redirect), is always a good thing. For example, changing #REDIRECT [[IBM Token ring]] to #REDIRECT [[IBM token ring]].
  • If an article links to a redirect because it has incorrect capitalization/etc, then cleaning that up without a pipe is a good idea. For example, changing [[IBM Token Ring]] to [[IBM token ring]].
  • Piping links that currently link to a disambiguation page is a good idea.
  • Changing links to redirects in real articles by just piping them (whether or not the current link text is correct) is often not a good idea – this is what wikipedians are often against. For example, changing [[Token Ring]] to [[IBM token ring|Token Ring]]
Hope that clears it up for you -- intgr 20:53, 9 December 2006 (UTC)
No problem. I should have left the redirects alone in the first place. In fact, I feel a little foolish about the whole thing. Anyway, good luck on settling this article under a title that everybody can agree on. It's an interesting entry on a stage in the development of LAN technology. Casey Abell 00:20, 11 December 2006 (UTC)
The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

"Burst" mode[edit]

Correct if I'm wrong, but I thought I remembered support for some kind of "burst" mode (may be the wrong terminology; I know that term is used with system bus architectures). Specifically, I thought there was a way where a node that had the token could submit multiple packets in rapid sequence. I didn't see anything about that in this article. EJSawyer 21:04, 13 March 2007 (UTC)

{Sorry to be so late joining this talk) I think you may be referring to "Early Token Release" - an enhancement that IBM introduced when it added 16 Mbps capability to the original 4 Mbps capability of the first Token Ring adapter cards. But it did not work in quite the way you suggest. In the original 4 Mbps Token Ring implementation, one 24-bit token normally circulated by itself (any "empty" transmission time ahead & behind being filled with idle 1-bits). When a station wanted to send a frame, it waited till the token arrived, and converted the token to a frame (by flipping a key bit in the second octet, inserting the data bytes of the frame immediately after, and then appending a final ending delimiter octet). This frame then "circulated the ring", by being passed from successive active station to active station until the frame came back to the sender. (NB. it was the SENDER that then removed the frame and replaced it by a new token. The RECEIVER did not modify the frame except to flip some bits at the back, in the so called Frame Status octet, to show the frame had been recognised.) By this mechanism the frame had to go right round the ring before the next token was released; and then that new token had to make it way right round the ring and back to the original sender before the latter could send its next frame. So multiple stations all wishing to send frames all got their chances to do so in strict rotation. When the speed was ramped up to 16 Mpbs this mechanism could lead to significant bandwidth being wasted if frames were small. The Early Token Release modification (default on virtually all newer adapters) solved this not by letting any station send more than one frame at a time (your "burst mode"), but by allowing multiple stations to send concurrently. This worked as follows. Each station wanting to send had to wait for the token as before, and as at 4 Mpbs (or non-ETR) the sender then substituted its frame for the token, but in ETR the new token was sent immediately after the end of that frame. Then any downstream station also wanting to send waited for the already-sent frame to pass and then replaced the token following it, by again substituting the frame to be sent and AGAIN appending a new token at the back. In other words, the general scene before ETR was that any time there was either just one token circulating, or one frame, but not both. With ETR the norm was a "train" of 0, 1, or more frames circulating, with a token always bringing up the rear. In theory every inserted station could have one frame in the train. Note that the order of the frames in the train always reflected the order of the sending stations around the ring. As before each sending station was responsible for removing the frame it had sent but now instead of there being just the one frame, there might be a train of multiple frames - however the sender's frame would always be leading when the train got back to the sender. It just had to strip the leading frame and let everything else go by, including the token bringing up the rear. Of course if it wanted to send another frame it would have to wait to put its new frame at the back of the train in the usual way. AconUK (talk) 16:18, 2 August 2014 (UTC)

Title should be "802.5 token ring"[edit]

The title makes it seem that this article is about token rings in general. That is not so; it is about 802.5 only. I don't see any discussion of other token rings, apart from a reference to the Cambridge ring. What about the token rings created by Apollo, or Proteon? (And there's FDDI, which is a very different protocol from 802.5.) Paul Koning 20:41, 16 April 2007 (UTC)

There are brief references to Apollo token ring and ProNet-10 in the overview... but that could be expanded quite a bit. Arguably, there should be separate articles for token ring (in general) and IBM/802.5 token ring, just as there are for bus network and Ethernet. --StuartBrady (Talk) 22:01, 16 April 2007 (UTC)


Token ring was develpoed by IBM for LAN. IEEE develpoed IEEE 802 project to standardize LAN. The IEEE 802.5 is almost identical to and completely compatible with IBM’s Token Ring network.Token Ring generally is used to refer to both IBM’s Token Ring network and IEEE 802.5 networks. Data rate , baseband signal , access method , encoding are both identical in IBM and IEEE. Topology , media are not specified in IEEE 802.5 . FDDI use dual ring , asynchronous and synchronous data frame , a different token frame format , has a similar frame format though not same

Rait 17:21, 19 May 2007 (UTC)

I hate to suggest this after all the merge debate, but I think there should be a very short article, maybe even just a disambiguation page, titled "Token Ring," with pointers from there to IBM/802.5, Cambridge, Apollo, Proteon, etc. Rees11 19:47, 19 August 2007 (UTC)

Article cleanup[edit]

I think this article would benefit from being cleaned up. There seems to be missing information, as if half ideas were adding but never completely finished. —Preceding unsigned comment added by 212.219.142.33 (talk) 16:44, 21 November 2007 (UTC)

I agree. The introductory paragraph is terrible. It doesn't at all describe what "token ring" means. It just talks bout its popularity. The introductory paragraph should briefly describe token ring.   –Justin Force 20:45, 1 July 2008 (UTC)

agree; further amplification required: a comparison to ethernet states "When token ring LANs were first introduced at 4 Mbit/s, there were widely circulated claims that they were superior to Ethernet,[6] but these claims were fiercely debated.[7][8]". but those citations are to a study of ethernet loading that doesnt even compare token ring, and a rambling user group argument. — Preceding unsigned comment added by Donquixote2u (talkcontribs) 21:38, 14 October 2012 (UTC)

Hi - I'm a Wikipedia newbie, and wireless networks have superseded most/all the old Ethernet Token ring squabbles, but years back I used to be an IBM specialist with token ring skills. If it is still worth updating this article in 2014 now, I'd be happy to help. As regards some of the naming discussions, the title was, strictly, "token-passing ring" but at IBM it was often abbreviated to TRN for "token ring network". The "ring" did not reflect physical wiring but rather the circuitry, with a back circuit in each hub. But there was no continuous ring of wire, a separate circuit was used by each active station to send signals to its downstream neighbour. In a separate comment above (Burst mode) I've tried to explain the token management, and in particular I've emphasised that the sender, not the receiver, replaced any sent frame with a token. But that was only before ETR arrived.AconUK (talk) 16:42, 2 August 2014 (UTC)

Misunderstanding the ring[edit]

I'm just going to revert a very nice picture that someone's added. Why? Because it implies that with two or therre host connected to a MAU/MSAU there is no ring. There is - as the diagram we've had here for a while shows clearly.

Sorry if this seems mean spirited (it's a nice pic I reverting), but this misconception is very common, and I think it's improtant for Wikipedia to help clarify thingsSnori (talk) 09:39, 30 December 2007 (UTC)

The ring is closed within the MAU if there is nothing connected to the Out and In ports. But if something is connected, the 'virtual' ring opens up and the MAU tries to deal with the surrounding as if it was a real ring (with cables and everything). 83.253.237.178 (talk) 15:35, 2 January 2008 (UTC)
So shall I put the picture back or... I think the picture shows why the protocol is called token ring at all, something that is not obvious for anyone who has never been in contact with it before. There is so many other cases in computer sience where words only are used metaphorically. Of course the image text shall say that the MAU creates a virtual ring if there is no physical ring, and can thereby work alone. Magnus Andersson 83.253.237.178 (talk) 04:38, 10 January 2008 (UTC)
Is the article better with the picture than without? It sounds like it. So put it back in, then you or others can improve things further either by editing the picture, or by adding text to clarify those corner cases if necessary. Paul Koning (talk) 12:03, 10 January 2008 (UTC)
There, more pictures :) Magnus Andersson 83.253.245.169 (talk) 16:20, 11 January 2008 (UTC)
That is, I added a single MAU network to avoid the misconception that there has to be a physical ring, I also added MAU photo that I found. Magnus Andersson 83.253.245.169 (talk) 23:38, 11 January 2008 (UTC)

Cabling[edit]

Category 4 cable claims this was the cable used for Token ring. If this is true, perhaps it should be mentioned in this article. --Kvng (talk) 13:14, 6 July 2010 (UTC)

Certainly the shielded twisted pair cable that token ring cards require have a different impedance than Cat 4 (which I believe is 100 Ohms like cat 3 or telco cable). See http://www.epanorama.net/documents/wiring/twistedpair.html Shjacks45 (talk) 04:46, 27 June 2013 (UTC)

History section?[edit]

I think it would make this article more readable if the Description section could be split into Description and History sections. This would help reader who are more interested in the history of the technology than the technical detail, similar to the article for the related Ethernet technology. — Preceding unsigned comment added by 203.167.250.228 (talk) 23:52, 4 June 2013 (UTC)

missing stuff[edit]

Some good technical but kind of messy, repeated items about the ring. Most token ring cards (in PCs) that I've encountered were DB9 (look at the picture), and yes if RX and TX were split out you could setup a few computers in an actual ring. Unlike Arcnet however, when a computer was turned off, the network card didn't automatically complete the circuit. MAUs or passive hubs would simply bypass computers that were off or disconnected. Active hubs allowed for "source routing" c.f. it.wikipedia.org/wiki/Source_routing_(token_ring)‎ or en.wikipedia.org/wiki/Source_route_bridging‎, which sent "unroutable" LLC packets to other network locations. Token Ring networking was SNA (IBM Systems Network Architecture) over DLC (Data Link Control). NETBeui was sometimes used for Applications over Token Ring. Note that most early network cards were 1.) interrupt based, 2.) simplex not duplex (hubs also), 3.) required software to assist the card firmware: 10baseT and 100baseT network cards didn't (couldn't) come close to theoretical bandwidth. CSMA also requires a "random" holdback time before resend, and the resends add to network traffic. I co-managed a 400 person call center, (we had HP 48 port managed hubs at the time) and noted that peak (saturated) network traffic was below 30% of 100mbit/sec (lower with more people on a segment). So, yes Token Ring is not as fast on lightly loaded networks but speed is superior on heavily overloaded networks. See system3x.com/as400_cabling/ re Twinax, it was not Token Ring but a serial device bus for text terminals and printers. Note the use of RJ11 patch panels. Shjacks45 (talk) 09:49, 27 June 2013 (UTC)