Talk:100 Gigabit Ethernet

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Computing / Networking (Rated C-class, Mid-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.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by Networking task force (marked as Mid-importance).

what is the advantage of 40/100 gigabit ethernet[edit]

over simply using link aggregation with. I'm guessing it's higher efficiancy due to a better distribution of data between links but I'd like to see it mentioned in the article given that most of the 40/100 gigabit ethernet standards simply seem to be based on multiples of the 10 gigabit stuff. Plugwash (talk) 22:12, 22 April 2009 (UTC)

it should be noted what the difference between 40 gigabit and 100 gigabit is. —Preceding unsigned comment added by (talk) 05:26, 18 July 2009 (UTC)
Must be a trick question. The answer is 60 gigabit. [sic]  :-) W Nowicki (talk) 20:00, 7 June 2011 (UTC)
Added a link to Dense wavelength-division multiplexing to the article header. -- (talk) 08:37, 24 August 2009 (UTC) —Preceding unsigned comment added by Simbad87 (talkcontribs) 15:56, 2 December 2009 (UTC)

merge IEEE 802.3ba[edit]

The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section. A summary of the conclusions reached follows.
The result was merge IEEE 802.3ba into 100 Gigabit Ethernet. -—Kvng 02:09, 14 January 2013 (UTC)


This can be merged with IEEE 802.3ba. `a5b (talk) 00:54, 1 December 2010 (UTC)

Which article should be the main page? Also, IEEE 802.3ba covers both 40 and 100 GbE - one standard covers two distinct application areas. Right now 40GbE redirects to 100GbE, which I find a bit odd. Opticalgirl (talk) 14:51, 9 December 2010 (UTC)
What can you add to the article "40GbE"? History of 802.3ba standard and link to it? This 802.3ba contains description of two speeds, so it can be main article. The 40GbE implementation is only 802.3ba at the moment. 100GbE's is too. Consider "40 and 100 Gigabit Ethernet" vs "IEEE 802.3ba". But we also can move the "802.3ba" article into this one (like 10 Gigabit Ethernet) `a5b (talk) 18:23, 9 December 2010 (UTC)
It would be a bad idea to make 802.3ba the main article. There will be other 40G and 100G Ethernet standards besides 802.3ba. For example 802.3bg is close to publication (40G serial) and a study group has just formed to look into 100G backplane. This is likely to lead to a new standard. Having separate articles for 40G and 100G can be justified. There are different applications for the two speeds, different products and different standards. —Preceding unsigned comment added by (talk) 16:32, 3 January 2011 (UTC)
Agreed that present PMDs are specific to 802.3ba, but it would still be useful to cite, and, as new PMDs are created, to reference the formulating IEEE document. This would provide a useful reference compendium. Not sure if any group other than IEEE can legally create an "Ethernet" standard. Until IEEE completely republishes 802.3 incorporating the work of 802.3ba, there is no better way of precisely referencing the work. Other Ethernet activities, even though re-published, are still referenced by the formulating committee; the utility is that it permits those interested to access the publicly-available work of the group. Fabrice002 (talk) 02:31, 24 June 2011 (UTC)
I do not follow the above. What exactly is your proposal? Clearly the standard needs to be cited; generally more citations are needed. Nobody propsed removing a citation of the standard. What I propose is to merge the IEEE 802.3ba into this one by combining the information and sources, leaving that as a redirect here. My reasoning is that IEEE 802.3ba is now "done". It will not be more notable since the group has finished. Other standards like bg are done in other task forces. And to answer your question: no, at least in the USA there is no law that give the IEEE a monopoly. Other groups can certainly and have in the past created other variants of Ethernet. In fact, there are several 40 Gigabit variants that vendors seems to have done. It would be better if this were clarified with sources cited in the article. W Nowicki (talk) 17:22, 24 June 2011 (UTC)

I sopport a merge. If we want to keep IEEE 802.3ba around, it should be about the task group, not about the technology. Whether to split 40 and 100 Gbit into separate articles is a separate issue. I think they can stay in the same article for a while longer until this thing sorts itself out. --Kvng (talk) 23:24, 23 January 2011 (UTC)

Yes, one article with proper citations and advertising removed, written in past tense please. Then add more later. W Nowicki (talk) 03:31, 5 June 2011 (UTC)

With the approval of 802.3bg as another PHY, I vote we merge the other way, and make this the resulting article. W Nowicki (talk) 20:00, 7 June 2011 (UTC)

The above discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.

Dead link[edit]

During several automated bot runs the following external link was found to be unavailable. Please check if the link is in fact down and fix or remove it in that case!

--JeffGBot (talk) 21:18, 31 May 2011 (UTC)

It looks like that was already archived? So just change the "deadlink=no" I suppose. The same info is still on the site, but they came out with more products, see this page.

But in general this should not be a buying guide. So most of those ads need to be removed, or at least edited to summarize. W Nowicki (talk) 20:00, 7 June 2011 (UTC)

Odd paragraph[edit]

There is, as yet, no commercially available PMD Layer 1 compliance test equipment for the 10 Gb/s multi-lane PMDs: 40GBASE-xR4 and 100GBASE-SR10. Most existing 40G and 100G test equipment is for verification of higher-level protocols. L1 capability is usually limited to bitstream BER, and cannot evaluate transceiver conformance to PMD optical and electrical specifications contained in IEEE 802.3ba, which includes eye mask, jitter tolerance, Transmitter and Dispersion Power Penalty (TDP), and Stressed Receiver Sensitivity (SRS), among others. Broad-based physical layer test capability may be prohibitively expensive due to the number of PMDs, and to differences between the unique methods created for 40- and 100GBASE- as compared to existing 10GBASE- methods, despite the same bit rate. PMDs differ, too, in wavelength (device technologies), number of lanes, fiber types and break-out methods. 100GBASE-LR4 and -ER4 test equipment, operating at 25 Gb/s, is not applicable to the 10 Gb/s PMDs. Absent commercially-available equipment for the shorter reach, lower-cost, and higher volume PMDs, users may revert to existing 10GBASE- test methods.

This was uncited and full of jargon. Might be something here, if we can dig up sources. Above was from IP address on June 4, 2011. W Nowicki (talk) 22:48, 7 June 2011 (UTC)

Odd comment...[edit]

... given that every term and and concept is used in the IEEE 802.3ba 40/100 GbE standard, or was used to describe capabilities of previously cited products. This jargon comprises specific compliance requirements of the IEEE standard, which themselves derive from normal practice by those skilled in the art of serial communication links. Citations of 802.3ba sections and paragraphs can be added, if required [how many, and how specific should they be?], but removal of factual content was unjustified. One might hope that technical vs. editorial/procedural expertise is not sufficient justification for censorship.

I will re-insert paragraph with minor edits shortly.Fabrice002 (talk) 02:18, 24 June 2011 (UTC)

Do not follow that comment either. The Wikipedia:Verifiability guidelines say "Anything that requires but lacks a source may be removed". Now since "you cannot prove a negative", statements like "There is, as yet, no commercially available..." cannot be proven, so therefore sound like personal opinions. Very dated ones at that, since if any such product does come out, the statement would become false. Such a statement might be true at the time you add it, but it is not encyclopedic. As for jargon, remember that people read these articles because they do not know the subject matter already. If they already knew it, they would not need to read the article, right? And if the statement is just generally giving engineering practice, then it does not seem specific to this technology. For example, there are multiple PMDs for 10 Gigabit Ethernet and Gigabit and every other speed too.
Ah, a line that was added by anonymous user at the same IP address might help explain it, I think. Maybe that the testing products available are intended to test switches and network interfaces, not the transceivers themselves? That might make sense, maybe as an intro to this section. After another quick look at the JDSU news release, it is till not clear to me what was being "introduced". It says "JDSU offers a complete solution for transponder characterization" but hard to tell if that is just a marketing brag. Anyway, the testing part is probably not the "meat" of the article since more readers might be more interested in how "real" the product is and where it tends to be in use. I will try to fill in some other citations. W Nowicki (talk) 19:34, 24 June 2011 (UTC)
You are correct regarding the statement of "previously cited products." I was referring to the list of Ethernet protocol test equipment that had been listed, several of which claimed "L1" capability. The intent of the statement regarding lack of availability of PMD compliance equipment was:
 - to stimulate anyone who can identify such equipment to state so, and improve the text;
 - to highlight limitations of the protocol test equipment laying claim to L1 capability, and
 - to provoke interest in the need for practical PMD compliance tests or development of surrogates.
I have searched unsuccessfully for this equipment for nearly a year, concentrating on manufacturers of prior generations. The industry is in need, but equipment manufacturers have privately said it is not forthcoming. Lack of this capability can create conflicts in attempts to use the standard, possibly impairing its adoption. The new standard deviates only slightly from existing 10G methods, but sufficiently to obsolete the equipment. At present, users of "Ethernet compliant" transceivers are unable to replicate, verify, or evaluate that claim in a defensible manner, that is, in accordance with the standard. I want to raise awareness of this situation.
If you would identify which statements "require" a source, I will locate. Fabrice002 (talk) 01:31, 27 June 2011 (UTC)

Conflict in Physical Standards section[edit]

A table at the beginning of the section lists the standard that applies to 7 meter copper cables and that is the only copper cable entry. Shortly after that table, it says: The 10 m copper cable objective is met with 4 or 10 differential lanes using SFF-8642 and SFF-8436 connectors. It seems that either the 7 m should be 10 m or the 10 m should be 7 m. Buchs (talk) 19:10, 2 April 2012 (UTC)

The source cited does not give any information about distances. I have removed both distance specifications. --Kvng (talk) 18:38, 4 April 2012 (UTC)

what does "40G" mean?[edit]

I cannot see "40G" explained anywhere. Does it need a link? Or perhaps someone can include an explanation in the article. Dondervogel 2 (talk) 22:02, 23 June 2013 (UTC)

It is just jargon. The people "in the know" use "40G" "40 gigabit" "40GbE" or other variants as shorthand contractions, since the meaning is usually known from context. In Wikipedia the convention is to spell out "40 Gbit/s" (officially with a non-breaking space between the number and units, but that makes the markup less readable). Since the lead explains the contractions "40GbE" I suppose we could use that in the rest of the article, although paraphrasing to "40 Gbit/s" also seems reasonable. But just "40G" seems like a bad idea. Especially due to the over-use of the marketing buzzword "4G" etc. Sources are very inconsistent. Almost all drop the "per second" since it is clearly a rate, using upper case G most of the time, but usually "b" not "bit", but I propose we use Wikipedia rules in the article. Perhaps explaining how the industry is inconsistent with the naming. I see there is a disambig page to 40G (disambiguation) which is fine, but like most disambig pages, any links to it should be changed to spell out the rate less ambiguously. Even then, the rates are somewhat approximate, since various overheads could be counted or not, so can be a little more or a little less that exactly 40 billion bits per second. In general this page still needs much work giving a better context in the lead, instead of jumping right into the alphabet (and numeric!) soup of jargon like IEEE 802.3az SDH PHY etc. W Nowicki (talk) 16:38, 24 June 2013 (UTC)
Thank you for this clear explanation. Your changes have made the article much clearer. Dondervogel 2 (talk) 19:32, 24 June 2013 (UTC)

Products section[edit]

Any objections to removing the products section? It's horribly out of date and would have to be updated weekly, almost, to keep it current. I would suggest either scrapping it entirely or replacing it with a (short!) section listing the most significant advances/releases. Also, there is much time-sensitive information elsewhere in the article that should be identified by date of writing - or perhaps removed instead. Madgenberyl (talk) 17:12, 6 April 2014 (UTC)

No objection here. ~KvnG 14:26, 10 April 2014 (UTC)

Wots in a name[edit]

Following discussion, 100 Gigabit Ethernet was unanimously accepted as the most correct name. Additionally, contributors agreed that other similar articles (specifically gigabit Ethernet and 10-gigabit Ethernet) should be renamed accordingly and consistently. Samsara 17:00, 7 December 2014 (UTC)
The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

I oppose the change of name. Leaving aside the hyphenation for the moment, I think the "Gigabit Ethernet" needs to be u/c "G" and u/c "E" because that is its name. "100-gigabit" is certainly not an accurate description of its bandwidth. Dondervogel 2 (talk) 15:18, 29 September 2014 (UTC)

I assume you are arguing that 100 Gigabit Ethernet is a proper noun and the official name for the technology? Do you have any evidence? I don't really have any evidence for or against. I made the name change here to be in line with what has been done at gigabit Ethernet and 10-gigabit Ethernet. ~KvnG 17:04, 29 September 2014 (UTC)
It looks like User:intgr and I were involved in naming 10-gigabit Ethernet. Happy to have more input on this from User:Dondervogel 2 and others. We could even do an WP:RFC on it. ~KvnG 18:59, 29 September 2014 (UTC)
Yes, I am arguing that 100 Gigabit Ethernet must be a proper noun, not because I have evidence to support it, but because the opposite assumption, that it as an Ethernet whose capacity is 100 Gbit, makes no sense (I assume that what is meant is a bandwidth of 100 Gbit/s). Sensible names of the article therefore include 100 gigabit per second Ethernet and 100 Gigabit Ethernet, but never 100 gigabit Ethernet. Dondervogel 2 (talk) 23:00, 29 September 2014 (UTC)
Good argument. I'd also tend towards 100 Gigabit Ethernet as a proper noun. Conquerist (talk) 00:26, 2 October 2014 (UTC)
Looks like the task force that developed this was thinking along the lines of your second suggestion: 100 Gbit/s Ethernet. Looks like most sources in a quick Google search go with a proper noun interpretation: 100 Gigabit Ethernet. Of these choices, I prefer 100 Gigabit Ethernet but past experience indicates that I'm pretty bad at making these calls. ~KvnG 02:16, 30 September 2014 (UTC)
I find it hard to get worked up about it either way, but I too prefer 100 Gigabit Ethernet to 100 gigabit Ethernet. Shall we wait a couple of days to see whether anyone else expresses an interest, and take it from there? Feel free to initiate an RfC if you prefer that. Dondervogel 2 (talk) 07:12, 30 September 2014 (UTC)

What name should this article have? I think a bit more discussion is required to resolve this. Naming affects not just article titles but many wikilinks and unlinked references. ~KvnG 01:29, 23 October 2014 (UTC)

My preference is "100 Gigabit Ethernet", for the reasons given above. Dondervogel 2 (talk) 10:14, 26 October 2014 (UTC)
"100 gigabit Ethernet", to follow the convention on similar articles:
I don't have an opinion on which convention is better, but it should be the same one for all the articles on specific Ethernet speeds.WarKosign 07:10, 9 November 2014 (UTC)
Be aware that some of these examples are my work and I'm no longer confident I did the right thing in those cases.
I'm aware of two objections to "100 gigabit Ethernet":
  1. "100 gigabit" is a compound modifier so grammar dictates a hyphen when it precedes a noun i.e. "100-gigabit Ethernet"
  2. "100 gigabit" is a quantity; we're trying to describe a rate. "100 gigabit/second" would be technically correct.
As I've come to see it, we either do something technically and grammatically correct i.e. "100-gigabit/second Ethernet". Or we declare it an informal proper noun or title and go by what people call it i.e. "100 Gigabit Ethernet" (Gigabit is capitalized here because each word in a proper noun is capitalized.) ~KvnG 19:05, 11 November 2014 (UTC)
I prefer 100 Gigabit Ethernet because it's shorter, but either will do as long we fix all the articles. WarKosign 19:13, 11 November 2014 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

@Dsimic: You've reverted a couple of my edits. Can you please explain why you believe the hyphenated forms are better ? I do not mind either way, but prefer all Ethernet articles to be consistent. WarKosign 04:40, 9 December 2014 (UTC)

Hello! Thank you for pointing out this discussion, but please let me read everything first. — Dsimic (talk | contribs) 06:02, 9 December 2014 (UTC)
Ok, 10 Megabit Ethernet, Gigabit Ethernet, 10 Gigabit Ethernet, and 100 Gigabit Ethernet they are – makes more sense than using 10-gigabit Ethernet, for example, which would sound as it's just one of some Ethernets, while it's a standard of its own. Out of those, 10 Megabit Ethernet could be a bit unsuitable, but let's sacrifice a bit of precision for consistency. :) Moving forward, I'll correct that when spotted in various articles, so we end up with broader consistency. — Dsimic (talk | contribs) 10:02, 9 December 2014 (UTC)
Thank you. 100-megabit is technically incorrect, it should be either 100-megabit/second for speed or 100 Megabit as a name. WarKosign 10:07, 9 December 2014 (UTC)
You're welcome. :) Right, "100-megabit" isn't a speed/bandwidth designation at all. — Dsimic (talk | contribs) 10:18, 9 December 2014 (UTC)

Thanks everyone for constructive discussion and resolution. ~KvnG 17:11, 9 December 2014 (UTC)

Ditto, with special thanks to KvnG who is going to the considerable trouble of implementing the ensuing name changes. Dondervogel 2 (talk) 18:01, 9 December 2014 (UTC)
Nice to have a persistent and reasonable scheme. Just for completeness, IEEE 802.3-2012 isn't too consistent by itself. Older standards are named "10 Mb/s Ethernet", "100 Mb/s Ethernet" but then they changed to "Gigabit Ethernet" and "10/40/100 Gigabit Ethernet". Good job! -- Zac67 (talk) 19:41, 9 December 2014 (UTC)
Right, it's great to have a consistent naming scheme we can apply all around. — Dsimic (talk | contribs) 21:56, 9 December 2014 (UTC)

support for twisted-pair cable at 40GbE speeds[edit]

40GBASE-T standard supports 40 Gbit/sec using twisted-pair cable, so you may want to edit "summarized port types" in lead. -- (talk) 06:12, 29 August 2016 (UTC)

Changed - thx. --Zac67 (talk) 06:28, 29 August 2016 (UTC)