karsh@geowhiz.UUCP (Bruce Karsh) (05/17/85)
We are trying to set up Ethernet here. We have 3 Masscomp systems
running the Excelan 201 board, and an Imagen laser printer running the
3-Com board. We are having a hard time trying to figure out which
transcievers and transciever cables to order.
As I understand it, there are 3 different Ethernet standards,
Ethernet 1, Ethernet 2, and IEEE 802.3. I have been told that it is OK
to mix and match transcievers for these standards on a single Ethernet
coax.
QUESTION 1: Is it ok to mix and match the various Ethernet
standard's transcievers on the same coax?
I have been told that you must pick a transciever cable that matches
the transciever you want to connect to. So if I want to connect to
an 802.3 board, I have to get an 802.3 cable. It doesn't work to
hook an 802.3 board to an Ethernet 2 board.
QUESTION 2: Does my transciever cable have to match the same standard
as my Ethernet interface?
QUESTION 3: What is the difference between the various transciever
cables? Why are they not all alike?
The TCL INC price list lists two kinds of transciever cables, Transciever
cables and 802.3 Transciever cables.
QUESTION 4: Which specs work with which cables?
QUESTION 5: Which standards does the 3-Com board comply with. Which
standard does the Excelan 201 comply with.
******************** FLAME ON *********************
I have been buying computer peripherals for seven years now. I have
never had this much trouble getting product information for *any* other
product. When you call Ethernet manufacturers, you invariably get some
empty headed jerk that knows absolutely nothing about the product, but
is perfectly willing to give you his totally uninformed advice on what
to do. When you push for a better answer, they contradict themselves.
I have been told so many contradictory things about Ethernet parts
that I can't believe anything I hear. The conflicting standards for
Ethernet are going to set LAN's back ten years.
We'd be better off with no standard at all.
(I've been on the net for about a year now. I've never flamed anything
before.)
******************** Flame Off **********************
(Boy, that sure felt good.)
I'd appreciate it if somebody could give me some help with this. If
somebody knows of some good references for more information, I'd
appreciate that too.
--
Bruce Karsh |
U. Wisc. Dept. Geology and Geophysics |
1215 W Dayton, Madison, WI 53706 | This space for rent.
(608) 262-1697 |
{ihnp4,seismo}!uwvax!geowhiz!karsh |hedrick@topaz.ARPA (Chuck Hedrick) (05/20/85)
In article <185@geowhiz.UUCP> karsh@geowhiz.UUCP (Bruce Karsh)
asks about the various Ethernet standards.
In our experience, the situation with Ethernet is not as dismal as it sounds
from your note. I understand why you are upset. No doubt few vendors can
really tell you which spec they really follow. Everyone these days claims
to support IEEE 802.3. I have concluded that this is because of a desire on
the part of the marketing people to look "industry-compatible," but that in
most cases this claim shouldn't be taken too seriously. Apparently
marketers have gotten the idea that IEEE 802.3 and Ethernet are the same
thing, so they just add IEEE to all their publicity.
First, Ethernet 1 vs. 2. This does not affect what finally gets on then
coax. The primary difference is the way the controller talks to the
transceiver. (There is also some sort of system-wide maintenance protocol
defined in Ethernet 2, but I haven't yet see anyone who supports that.) As
long as your system can communicate with its transceiver, you don't have to
worry about whether it is using Ethernet 1 or 2. Indeed many of the newer
equipment is designed to work either way. So it is hard to categorize
things cleanly as one or the other. In practice, the way you go about
putting together a system is not by asking which standard the pieces follow.
Rather, you start with a system and ask what transceivers it can talk to.
While your salesman may not be sure whether his equipment is Ethernet 1 or
2, he almost certainly can give you a list of transceivers that have been
used successfully with it. This is really an empirical question anyway,
since there are some pairs of products which should talk and can't, and visa
versa.
As for IEEE 802.3. This standard is a problem. As far as I can tell (as a
non-expert), the change they have made has no advantage. (Indeed in most
cases, it causes the software to become noticably more complex.) You should
not blame this one on the Ethernet community. It is a typical standards
committee product. It is not just an issue of talking to the transceiver
this time. There is an actual difference in the interpretation of the
Ethernet header. If anyone actually implemented this standard as described
in the documents, he would be unable to talk to any of the standard TCP/IP
implementations, which are Ethernet-based. Fortunately, it is possible to
implement IEEE 802.3 in a way that maintains compatibility with machines
that use the Ethernet definitions. I don't know whether any vendors have
actually implemented IEEE 802.3, and if so whether they did the extra work
needed to be compatible with Ethernet. So far we haven't run into any real
implementations of IEEE 802.3. (As I said above, most of the people who
mention IEEE 802.3 don't really mean it.) The simplest way to determine
whether your system is going to have trouble is to ask the salesman whether
it can talk to a VAX running 4.2BSD. This cuts through the morass of
definitions and gets down to the issue you are really interested in: whether
it is compatible with the Ethernet spec. If the answer is no, no change in
transceivers or cables will help you. The difference involves the software
device driver. I can be fairly sure that the Imagen laser printer is
compatible with Ethernet. I'm sure they are not about to bring out a
product that won't talk to a VAX. Masscomp has had TCP/IP for a number of
years, so theirs is probably also Ethernet-compatible. I suspect the only
case where you have to worry is with a vendor that has no contact with the
Unix community, and is implementing TCP/IP from scratch using the protocol
documents. (Believe it or not, this does happen.) Neither of these vendors
falls into this category.
You asked whether there is any problem putting these various things on the
same cable. The only problem I know of is if one of the systems really does
implement IEEE 802.3, and you also have systems that use the old PUP
protocols. As long as you are using TCP/IP, DECnet, and XNS, there should
be no problem have different things on the same cable.
We use the same kind of transceiver cable for all of our systems, whether
Ethernet 1 or 2. However if you are a relatively unsophisticated shop, the
simplest thing to do is to get the cable from the same people you get your
transceiver or system from. They can then guarantee compatibility.mccallum@opus.UUCP (Doug McCallum) (05/24/85)
> implement IEEE 802.3 in a way that maintains compatibility with machines > that use the Ethernet definitions. I don't know whether any vendors have > actually implemented IEEE 802.3, and if so whether they did the extra work > needed to be compatible with Ethernet. So far we haven't run into any real > implementations of IEEE 802.3. (As I said above, most of the people who Don't confuse the IEEE 802.3 standard with the IEEE 802.2 standard. 802.3 specifies a "media access method" of CSMA/CD and describes what is basically Ethernet Version 2.0 with the type field called a length field. It is possible to be IEEE 802.3 compatible without implementing the IEEE 802.2 "Logical Link Control" protocol. The vendors claiming 802.3 compatibity probably are. Full IEEE 802 compliance is another matter. Very few 802.2 implementations are available. Doug McCallum NBI, Inc. {ucbvax, allegra, ut-sally}!nbires!mccallum
phil@amdcad.UUCP (Phil Ngai) (05/29/85)
In article <1214@opus.UUCP> mccallum@opus.UUCP (Doug McCallum) writes: >"Logical Link Control" protocol. The vendors claiming 802.3 compatibity >probably are. Full IEEE 802 compliance is another matter. Very few >802.2 implementations are available. Speaking of standards compliance, I hear a nasty rumor that some NBI product (workstation?) implements a RS-232 connector with pins 2 and 3 as DTE and pin 6 as DCE. Never mind that the damn thing is female too. (according to the standard, all DTEs, which includes terminals and computers, are to have a male connector) I've seen some wrong stuff before but this is a new record, if true. Please, say it isn't so. -- What do you do the day after a peak experience? Phil Ngai (408) 749-5720 UUCP: {ucbvax,decwrl,ihnp4,allegra}!amdcad!phil ARPA: amdcad!phil@decwrl.ARPA