ethernet hub

You've been told!

You don't seem to think of marketing as legitimate. And so are insulting Rich. I don't much like ad-hominem, and feel compelled to defend to my meagre ability:

I don't think _you_ understand what Ethernet actually is. Perhaps it will help to understand what Ethernet is _not_: It is not Token Ring (802.5) or Token Bus (802.4). It is not wireless (802.11). See

formatting link
As I understand it, Ethernet is the IEEE standard for CSMA/CD. This idea is very powerful and enables attractive technologies. There are many implementations: 802.3_ (10base5) 802.3a (10base2)

802.3b (10broad36) and so on. All up until GBE (802.3ae) allowed shared media, at first physically the long yellow garden hose, and later shared electricly. 100baseTX repeater hubs used to be made.

Now, it might have been said that GBE was a departure from CSMA/CD and really ought to have started a new committee, say 802.13. I'm sure it was discussed and I accept the learned decision. I believe that some of the essential features of CSMA/CD (gaps) have greatly simplified the switching electronics which might be said to be implementing a "virtual" CSMA/CD.

Perhaps you should be grateful 802.11 isn't also Ethernet! After all, it does CSMA/CD and 802.11b might be called 1broad1 through a medium of low conductivity :)

-- Robert

Reply to
Robert Redelmeier
Loading thread data ...

I didn't say I reject his claim. Only that I am questioning the arguments behind it. He is still alive, so is R Seifert. It's not illegal.

I don't think i'm insulting Rich, I read many of his posts and own one of his books. What greater praise can there be for an author/inventor/writer? Not all want to be treated as deities.

I know. It appears you want to insult me because you feel I insulted your deity. I did not.

R Seifert has explained the marketting definition of ethernet better. No need for you to

you can't have your cake and eat it. Meaning that 802.3x (and I guess GBE) aren't CSMA/CD. You can't take both definitions - both sentences - seriously.

It doesn't. it does CSMA/CA

And I didn't insult R Seifert, who I respect as an inventor, and one of only a few people that founded something, with enormous technical expertise, takes the care to clarify things on usenet.

Reply to
q_q_anonymous

q_q snipped-for-privacy@yahoo.co.uk wrote in part:

I am sorry but I do not see that. You have asked only three questions in this subthread, all rhetorical. You seem to be disagreeing, not questioning. And your disagreement has been far louder than your basis for it [full duplex].

But in directing arguments personally, you are being impolite. Most scientific discourse happens in the passive or third person to make it clear the ideas are being disputed, not the person.

Many things: respect and being positively influenced by their ideas. Do you never read/buy authors you do not wish to praise?

Fully agreed. I'm not treating Rich as a deity. I've stood up for alleged trolls and other undesireables when they have been unjustifiably attacked.

I do not see two definitions here, or any inconsistancy between my two sentences. You seem to think full duplex (802.3x) is not Ethernet because there is no possibility of collision. But there _are_ on the many networks with defective wiring (split pairs). Crosstalk mangles data almost as effectively as a collision. And the CSMA/CD vestigle features handle it.

Really? How? I'm not that familiar. Collisions have been a feature of transmission for 100+ years [radio]. Any medium with delay is subject to them. For most of that time, central direction has been the chief collision technology. The genius of CSMA/CD is to expect and allow collisions by dealing with them gracefully. Does 802.11 use some form of effectively central direction?

-- Robert

Reply to
Robert Redelmeier

"q_q snipped-for-privacy@yahoo.co.uk" wrote: [long rant]

What's your point? If I goto the store and ask for a 24-port 10/100 Ethernet switch, the clerk isn't going to refuse to sell it to me because it does full-duplex. The specs are all about compatabilty, not about ensuring that everyone is using exactly correct terminology. Everyone (else) agrees on what Ethernet is and means and how it works together, if you want absolute precision learn to program. 8*)

Reply to
William P.N. Smith

I wrote a long post to make all reasons clear. Not to shout. Infact, my last paragraph was a bit humouress.

[full duplex] is very relevant , and there were other reasons and issues, you can't summarise my post into [full duplex].

, they wanted a different dot, but for political reasons, it was included in 802.3, as if it were a mode of 802.3 So i'm not disagreeing with R.S. like you think, the arguments there rest in my reasoning with such facts. I am merely questioning his position on the issue.

are you serious?

if I didn't like his book then it'd be thrown out or in a dusty corner.

I bought his book, I read it, it's on my shelf. I quoted from it. I'm sorry if that isn't sufficient praise for your deity.

My knowledge is from reading some books, not much in electronics. I said that 802.3x has no CSMA/CD. That's almost the limit of my knowledge on 802.3x. I was basing that on an expert, the man himself, chair of 802.3x. I'll quote him.

"Full-duplex mode eliminates the entire CSMA/CD algorithm: no collision detect, no jam, no backoff, no retransmission".

How 'crosstalk' is dealt with is another matter(with my limted background, i'm nto even familiar with the concept), but it's not a collision, and it isn't treated like one.

I guess 802.3x has no collisons.

I studied a bit of 802.3. I haven't studied 802.11, haven't 802.3x either. My interest is in networking generally , i'm at Layer 3 of the OSI . When I reach Layer 7 i plan to come back down and have a look at some other LAN technologies - layers1,2.

I hope others can respond to your query. One post which I did read on

802.11 a while back, which might help, is this one

"802.11 listen for collision while transmitting?"

According to that thread, all there agree, there are collisions with

802.11
Reply to
q_q_anonymous

if you're not interested in precision, stay away from programming.

I can also talk in marketting terms, i'd do so if I have to talk to a clerk that only understands that. But i'm not going to talk to everybody as if they're that clerk.

Aside from that, I'm glad that you and a clerk can understand each other.

Reply to
q_q_anonymous

"Technical meaning" can better be described as a "term of art;" a word is a term of art (i.e., it has a "technical meaning") if the word implies something specific to persons skilled in the field of that art (and may mean something different to persons unskilled in the art). The meaning of such a term can change over time, as the technology evolves. For example, the meaning of "computer" to a skilled person in 1965 would not be identical to the meaning of "computer" to a skilled person today.

When it was first commercially promulgated in 1980, "Ethernet" meant a coaxial-cable based, CSMA/CD, half-duplex LAN operating at 10 Mb/s. The technology has evolved, and has been adopted and incorporated into IEEE

802.3. Today, "Ethernet" includes many data rates, full- and half-duplex modes of operation, many media, etc. Most people skilled in the art consider "Ethernet" to mean any system that conforms to any of the IEEE 802.3 family of standards, and often any component of such a system (e.g., an "Ethernet cable" or an "Ethernet chip"). It's not so much the technical people sacrificing their standards (pun intended) for marketing purposes, but the terminology itself changing as the technology advances.

I have said many times that the only thing that "today's Ethernet" has in common with "the original Ethernet" is the frame format (and even that gets tweaked now and then, e.g., VLAN extensions). That doesn't make it any less "Ethernet."

[snip]

If it behaves like Ethernet (to the client protocols and applications), and it exchanges Ethernet frames with its peer entity on the communications channel, why is it not proper to call it "Ethernet"?

The only thing that changed in 802.3x is the MAC algorithm. The MAC algorithm is a function of the underlying physical channel, i.e., the algorithm is *designed* to take into account the behavior and limitations of the channel. On a shared-medium channel, one needs an arbitration mechanism so that each station can determine when to send its frames without interfering with other stations. On a dedicated-medium channel, there is no such requirement. When Ethernet is used on a shared medium, CSMA/CD is the algorithm used to arbitrate for the channel. When Ethernet is used on dedicated media (as is required for full-duplex operation), there is no need to burden the station with the arbitration algorithm, so we dropped it, resulting in "full-duplex Ethernet" (IEEE 802.3x).

Would you have preferred that we kept the CSMA/CD algorithm in "for old times sake" even though it would degrade performance? Would you have preferred that we invent an entirely new LAN technology, with a new frame format (and called it something other than Ethernet) to take advantage of dedicated media? What we did was extend Ethernet concepts (frame format, connectionless exchange, transceiver and cabling technology, common client/driver interfaces, etc.) to the dedicated medium environment, and jettisoned all of the baggage that was unnecessary in that realm. That's the right thing to do from a

*technical* perspective. What benefit would have derived from calling it something other than what it was--Full Duplex Ethernet, i.e., "Ethernet" operating on dedicated media, in as efficient a manner as possible?

Thank you for the kind words. I keep telling my girlfriend that I am a "god" but her perspective may be different from yours.

"Ethernet" has value both as a technical term, and as a marketing term.

As long as you understand how your *listener* is interpreting it, you can use the term properly in many contexts. When I speak to technology wizards, I recognize that they understand "Ethernet" to mean the realm of technologies encompassed by IEEE 802.3 (and even some variants thereof). When I speak to marketing people, I recognize that they understand "Ethernet" to mean that set of products that can be sold to someone under the "Ethernet" name.

Indeed, many people call IEEE 802.11 "Wireless Ethernet". While this surely flies in the face of any technology argument (it doesn't even use the same frame format as IEEE 802.3), I can live with it *when speaking to people to whom the technology difference is unimportant*. To many lay users, "Ethernet" means "LAN"--it is the only LAN they have ever known. So a wireless "LAN" that behaves (from the user and application perspective) exactly the same as their traditional Ethernet becomes a "Wireless Ethernet." I might get upset if the 802.11 folks started using that term within their standard, but it doesn't bother me when it is used by law students connecting to a campus LAN.

By your argument, we should stop calling them Microwave *Ovens*, because an oven is something that gets *hot*, which a microwave oven does not. We call them "ovens" because we *use them like ovens*, even if the technology has changed over time.

A product development engineer who is not involved with marketing is useless. "Engineering" is the art of solving customer problems with technology solutions, at a price that yields a profit. That requires both technology (for the solution) and marketing (to define the problem and determine the proper price).

A person who develops technology without regard to the marketplace is a

*scientist*, not an engineer.

Most successful engineers and entrepreneurs wear many hats, *all at the same time*. That's what makes us successful; we consider the technology, the market, manufacturing, ease of service and maintenance, etc., in balancing the conflicting pulls from each factor. If you can achieve the right balance, you get a winning product.

Again, you seem to think that "techies" are somehow separate from, and must use a lexicon different from, the rest of the world. To me, that is a very narrow view. I deal with technologists, scientists, marketers, customers, lawyers, and "dumb users"; I try to use language that is appropriate for each group, i.e., I take into consideration how the listener will understand what I am saying. There is no reason to invent different words for each group; I can use the same term (such as "Ethernet") in various contexts, as long as I know what my listener will think when I say it.

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

Reply to
Rich Seifert

The problem with wireless is that is it difficult (if not impossible in most situations) to *detect* that a collision occurred, since the received signal is so much weaker than the transmitted signal as seen at the transmitter. It's like trying to hear someone across the room whispering while you are screaming at them.

Thus, many shared-LAN wireless systems use a form of CSMA/CA (collision avoidance, rather than detection). There are lots of variants on the methods, including both 1-persistent CSMA (i.e., backoff on carrier sense, without detecting collisions), RTS/CTS schemes (where a very short, acknowledged frame is sent to acquire the channel before sending a long frame), and centrally-arbitrated schemes (which are possible with

802.11 access points as the arbitrator).

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

Reply to
Rich Seifert

yeah, I hadn't considered the similarity of common frame format. That is the soul of Ethernet. That does indeed make sense from a technical standpoint, and allow for the term to evolve technically, as you say, with extensions e.g. VLAN.

But, a minor point that doesn't change your argument - your use of the terms client protocols and applications. Call me a mug for the OSI reference model, but, client protocols are application layer aren't they? They wouldn't be aware of ethernet.

agreed. I'm glad that the more modern definition still has meaning in a technical context too and in the context of product development, which is really going to leave technical "fingerprints".

oh, it's the law students, I knew it! Those arrogant - not strictly logical - know-it-alls. Those busybodies. Those ..

Funnily enough, I see how from a strict product development context, even 802.11 was based on the idea of Ethernet . Though I suppose it still doesn't merit to be called ethernet in that context, since 802.3 still exists. It hasn't obsoleted yet. It's not the natural successor. Infact, I recall reading that Bob Metcalfe wasn't much of a fan of

802.11 , and you don't want it referred to as Ethernet either.

But. From a product development context- If microwave ovens are ovens. Isn't 802.11 ethernet? what with 802.11's collisions and the ideas it was based on.

yep. I hadn't considered the frame format thing. That gives calling

802.3 ethernet a technical basis, and a basis for the technical definition to evolve.

Given that. I can't see any diff in the way marketting currently use the term or the way technical have allowed the term to evolve. I suppose that - in the case of the term Ethernet - thus far - the right compromises have been made by either side

As long as marketting don't break the technical rules. i.e. the technical side can accomodate them, then it's fine. If there is a clash. Then there would be diff meanings(definitions) for diff contexts. One meaning shouldn't completely trump another.

perhaps it's safer if marketting just invented their own terms, like 'switch' !

lawyers/"dumb users"/ , when they talk about these things, should be understood but not obeyed. I guess marketters have some authority. But when terms clash between marketting and technical, that authority is very weak if the technical side don't make compromises to give their terms any legitimacy. I suppose that similarly, lawyers/dumb users have influence in driving the market, but they can only influence marketting and marketting can only influence technical.

Thanks for your comprehensive response.

Reply to
q_q_anonymous

The Network layer is a "client" of an Ethernet Data Link. Thus, IP is a "client protocol", which is aware of Ethernet at the service interface to the Ethernet Data Link. That is why an IP implementation does not submit more than 1500 bytes of data for transmission in any given frame.

Also, if you are such an "OSI mug," you realize that entities at higher layers (e.g., the Application layer or above) *can* interface directly to a Data Link entity such as Ethernet. When doing so, they: (1) forego any services provided by the "skipped" layers, and (2) have visibility into the nature of the Data Link being used, to the extent of its effect on the service interface. So, an Application layer protocol (e.g., DEC LAT) running over Ethernet would be very much aware of the fact that it was running on an Ethernet.

You *do* realize that I spent the last four years of my life in law school, as a law student? (I had my last final exam on Tuesday; the graduation ceremony is next week, but I can now add "J.D." to my list of academic degrees.)

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

Reply to
Rich Seifert

Obviously you have not read the fine print in the patriot act.

but you're prolly safe, being on the other side of the pond and all...

;-)

Reply to
snertking

Heaven help us.

Do you plan on taking the bar exam? If so, what state? CA?

What exectly are you planning on doing with the degree?

Reply to
snertking

I will sit for the California Bar at the end of July. While I don't plan to "practice law for a living," I have been doing a lot of work lately in the area of privacy law as applied to electronic communications systems (combining of my legal expertise, communications expertise, and a strong libertarian bent). In particular, I am working on the design of vehicular communications systems, with an eye towards making information privacy inherent in the communications protocols and message exchanges (e.g., anonymity guarantees where appropriate, encryption, authentication, etc.)

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

Reply to
Rich Seifert

When I'm programming, I'm as precise as {practical,possible}.

Do you talk about {Kibi,Mibi,Gibi}bytes as well? Do you rant about the storage capacity of the hard drives you buy?

How much RAM is in your computer?

Does anal retentive have a hyphen?

Reply to
William P.N. Smith

No, but I did give it a few minutes thought during the few days that it was 'the news'. I didn't reject it straight away thinking 'who cares'.

I think it wasn't a bad idea to make an attempt at dealing with the problem. But wasn't necessary, because the only people that mean 'decimal' in the context of {K,M,G} Bytes , are those marketting hard drives. So, if talking advertised HDD capacities, it is best to use that language. Furthermore, ...bibyte sounds bad. But, most seriously, it tried to run against the established definition , etched in stone, of KB,MB,GB . Unacceptable.

I build mine, but I don't pay attention to the spec. All comps are fast nowadays. But if I did care about spec, and had to think or talk about it, I'd use the established practice , which is that HDDs are an exception in being specified in mathematical {K,M,G}Bytes.

I don't know, because I don't use that expression, since it has the word 'anal' in it, it's often inappropriate. Also, I don't want to promote Freud.

If I were to use the term, then I would check to see if it has a hiphen or not.

So the answer to all your sasrcastic questions, were yes and no. Or rather, "no and yes". "no I don't" "but, yes , I take it seriously". i.e. My 'no' is for a good reason.

Unlike yourself, for whom your 'no' is for lack of caring. If I were as liberal with my with my application of terms as you, i'd say that you, in not caring about these issues, would be anal-expulsive. Sounds worse than anal-retentive, doesn't it?

Reply to
q_q_anonymous

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.