Gigabit ethernet

Gigabit ethernet

Post by Darrell Amundso » Fri, 26 Jun 1998 04:00:00



Could any one help me with gigabit ethernet. I'am a student in a
technical College looking
for information for a report. Can you tell me or point me in the
direction to find out if Gigabit
ethernet can used over copper rather than fiber.

 
 
 

Gigabit ethernet

Post by Craig Wiesne » Fri, 26 Jun 1998 04:00:00


Darrell,

You can find lots of information about Gigabit Ethernet at
http://www.gigabit-ethernet.org/
http://www.ccrc.wustl.edu/pub/ieee-tcgn/tcgn.html

In Rich Seifert's book on Gigabit Ethernet, which I highly
recommend, does cover 1000BASE-CX for Shielded
Twisted Pair (dual, 150ohm STP with a maximum useful
length of 25m for equipment interconnection within the
wiring closet). Pages 243-245 of Rich's book cover
the cabling in detail.

I don't think you'll be seeing Gigabit Ethernet to the
desktop over anything other than fiber.

If you'd like to order a copy of Rich's book and help
feed the homeless, please visit our web site's Reference
Library where you can read about the book and click
through to Amazon and buy it. If you do, we donate all
our proceeds to the Ecumenical Hunger Program and
the Second Harvest Food Bank. The library is at
http://www.wkmn.com/refer.html

Best of luck learning!

Craig Wiesner - WKMN Training - InternetWorking Education Specialists
http://www.wkmn.com/vend.html     Visit the "Faster Ethernet" vendor
list!!


> Could any one help me with gigabit ethernet. I'am a student in a
> technical College looking
> for information for a report. Can you tell me or point me in the
> direction to find out if Gigabit
> ethernet can used over copper rather than fiber.


 
 
 

Gigabit ethernet

Post by John Lundgre » Mon, 29 Jun 1998 04:00:00



> Darrell,
> You can find lots of information about Gigabit Ethernet at
> http://www.gigabit-ethernet.org/
> http://www.ccrc.wustl.edu/pub/ieee-tcgn/tcgn.html
> In Rich Seifert's book on Gigabit Ethernet, which I highly
> recommend, does cover 1000BASE-CX for Shielded
> Twisted Pair (dual, 150ohm STP with a maximum useful
> length of 25m for equipment interconnection within the
> wiring closet). Pages 243-245 of Rich's book cover
> the cabling in detail.
> I don't think you'll be seeing Gigabit Ethernet to the
> desktop over anything other than fiber.

That sounds scary.  Of course, a whole lot of people said years ago that
it's only a matter of time before fiber gets to the desktop.  Looks like
the time is rapidly approaching.  Maybe we should start a new motto:
"Fiber to the desktop by 2000".

Recently I saw some site with info that the cat5+ and cat6 were going to
be added soon to the 568 specs.  Strange..  We buy a piece of test
equipment and it becomes obsolete within a matter of months.  And all
that wiring having to be replaced...  

Re: Amazon, below.  Right now, Spamazon has such a bad reputation for
junk email that if I were you I would disassociate myself from such a
crassly commercial for-profit at everyone elses' expense company as
Spamazon.  IMHO.  See news.admin.net-abuse.email newsgroup for more
info.

> If you'd like to order a copy of Rich's book and help
> feed the homeless, please visit our web site's Reference
> Library where you can read about the book and click
> through to Amazon and buy it. If you do, we donate all
> our proceeds to the Ecumenical Hunger Program and
> the Second Harvest Food Bank. The library is at
> http://www.wkmn.com/refer.html
> Best of luck learning!
> Craig Wiesner

> > Could any one help me with gigabit ethernet. I'am a student in a
> > technical College looking
> > for information for a report. Can you tell me or point me in the
> > direction to find out if Gigabit
> > ethernet can used over copper rather than fiber.

--
Please remove NO and SPAM from my email addr to reply.

Read what the FTC has to say about Junk Email at URL
http://www.ftc.gov/bcp/conline/pubs/online/inbox.htm
 
 
 

Gigabit ethernet

Post by Henry Spenc » Mon, 29 Jun 1998 04:00:00




Quote:>> I don't think you'll be seeing Gigabit Ethernet to the
>> desktop over anything other than fiber.

>That sounds scary.  Of course, a whole lot of people said years ago that
>it's only a matter of time before fiber gets to the desktop.  Looks like
>the time is rapidly approaching.

Maybe.  I remember hearing the same predictions made when the new hot
technology was 100Mbps.  Many people were sure that 100Mbps would never
make it to the desktop except over fiber.

Anyone trying to do Gbps over copper certainly has a job on their hands...
but I'm not prepared to count it out yet.  In the long run, there is a
*lot* of money waiting for someone who finds a way to do 1000Mbps to the
desktop *without* wholesale recabling.  That sort of thing inspires truly
remarkable creativity sometimes.

One has to be very careful about "proofs" that such things are impossible.
Such problems are seldom resolved by frontal *; usually what happens
is that some clever lad notices a loophole in a commonly-held assumption.
--


 
 
 

Gigabit ethernet

Post by Frank X.Ma » Tue, 30 Jun 1998 04:00:00


On Sun, 28 Jun 1998 04:27:39 -0700, John Lundgren



>> Darrell,

>> You can find lots of information about Gigabit Ethernet at
>> http://www.gigabit-ethernet.org/
>> http://www.ccrc.wustl.edu/pub/ieee-tcgn/tcgn.html

>> In Rich Seifert's book on Gigabit Ethernet, which I highly
>> recommend, does cover 1000BASE-CX for Shielded
>> Twisted Pair (dual, 150ohm STP with a maximum useful
>> length of 25m for equipment interconnection within the
>> wiring closet). Pages 243-245 of Rich's book cover
>> the cabling in detail.

>> I don't think you'll be seeing Gigabit Ethernet to the
>> desktop over anything other than fiber.

>That sounds scary.  Of course, a whole lot of people said years ago that
>it's only a matter of time before fiber gets to the desktop.  Looks like
>the time is rapidly approaching.  Maybe we should start a new motto:
>"Fiber to the desktop by 2000".

>Recently I saw some site with info that the cat5+ and cat6 were going to
>be added soon to the 568 specs.  Strange..  We buy a piece of test
>equipment and it becomes obsolete within a matter of months.  And all
>that wiring having to be replaced...  

It might interest the group to know that in some initial test on
transmission for gigabit ethernet that Cat 5 was more successful than
62.5/125 micron multimode fiber. There still isn't any application
that would necessitate the use of fiber to the desktop. The fiber
people have been promoting this since Hector was a pup. The latest
marketing flare-up is to try to sell you on their brand of the new
not-yet-a-standard type connector.
 
 
 

Gigabit ethernet

Post by john mos » Tue, 30 Jun 1998 04:00:00



>It might interest the group to know that in some initial test on
>transmission for gigabit ethernet that Cat 5 was more successful than
>62.5/125 micron multimode fiber.

I would be very interested to read about these results, engineering
data and specific conditions of the test. Where can I find these
results? What is the cost of the electronics? Are there distance
limitations? What is the status of the standard?

Quote:>There still isn't any application
>that would necessitate the use of fiber to the desktop.

Well, less than a decade ago 4 and 16 Mbs was plenty! Go figure.

Quote:>The fiber
>people have been promoting this since Hector was a pup.

Pardon my ignorance, who is Hector?

Quote:>The latest
>marketing flare-up is to try to sell you on their brand of the new
>not-yet-a-standard type connector.

Geez, CAT 6 and CAT 7 , no standard, and marketing?  Please!

It all boils down to cost,  which technique solves the problem and is
based on usable standards. I think cost will drive the market as
always. Let's wait and see before we blast away at innovation.

John

 
 
 

Gigabit ethernet

Post by Arne Kelle » Wed, 01 Jul 1998 04:00:00




>>It might interest the group to know that in some initial test on
>>transmission for gigabit ethernet that Cat 5 was more successful than
>>62.5/125 micron multimode fiber.

>I would be very interested to read about these results, engineering
>data and specific conditions of the test. Where can I find these
>results? What is the cost of the electronics? Are there distance
>limitations? What is the status of the standard?

Lucent has made a lot of testing on their new GigaSpeed cablesystem and they
have 1Gethernet like signals running on it.
And  Lucent garanties that 1Gethernet on copper vill run on GigaSpeed when
all the standards are ready.

Arne Keller

 
 
 

Gigabit ethernet

Post by John Row » Wed, 01 Jul 1998 04:00:00



> Maybe.  I remember hearing the same predictions made when the new hot
> technology was 100Mbps.  Many people were sure that 100Mbps would never
> make it to the desktop except over fiber.

I'm a new boy in this game but I can read, add up and even
multiply. Gigabit sounds like ten times as much work as 100BASE-TX but
it isn't: gigbit will use all four pairs and they plan to use full
duplex _on each pair_. If can they do it this gives them a factor of
four already. (And that's not even counting 100BASE-T2 - does that
really exist?)

Also, look at the 100BASE history. 100BASE-T4 allows transmission over
catagory 3 but I would guess that most people are using 100BASE-TX
over catagory 5. Why? Because anybody with an eye to the future was
installing catagory 5 when they only needed catagory 3. And whilst
it's _nice_ to have something that anybody can run, all you _need_ is
something that a fair number of people can run.  I'm going to be
installing 350 MHz cable everywhere and I guess that everybody else
with any sense is too.

One thing's for sure - if they can't get gigabit to run at 100 MHz
they certainly will be able to at 2-300 MHz. It wouldn't surprise me
at all if even if they do manage it at 100 MHz there won't also be a
1000BASE-TX version running at a higher frequency and requiring much
less electronics and that that's the version that will take off.

I reckon gigabit will run over high spec UTP, maybe even over vanilla
catagory 5. But we'll see.

John

 
 
 

Gigabit ethernet

Post by John Lundgre » Thu, 02 Jul 1998 04:00:00





> >>It might interest the group to know that in some initial test on
> >>transmission for gigabit ethernet that Cat 5 was more successful than
> >>62.5/125 micron multimode fiber.

I've heard something here that 62.5/125 multimode cable cable isn't good
for Gigabit speeds.  But then copper wasn't supposed to go that fast,
either.  The copper problem was gotten around by using all four pairs,
and sending and receiving on all pairs, so the actual speeds go up to
something like 250 MHZ.  

It seems that a design using something similar would work well on
fiber.  Like using wavelength division, where there are two LEDs or
LASERs to transmit, and two receivers for different wavelengths.  Sure,
this means that the system is more expensive, but then that would be
expected for a 10X increase in speed.  And as time goes on and the
volume increases, the cost would go down due to mass production.

Right now, I see info that the long haul fiber circuits are going up to
80 GBS using this wavelength muxing.  Of course they're on SM fiber, and
the cost of a hundred thousand dollar piece of equipment is a whole lot
cheaper than putting in miles of additional fiber.  :-)

Quote:> >I would be very interested to read about these results, engineering
> >data and specific conditions of the test. Where can I find these
> >results? What is the cost of the electronics? Are there distance
> >limitations? What is the status of the standard?

> Lucent has made a lot of testing on their new GigaSpeed cablesystem and they
> have 1Gethernet like signals running on it.
> And  Lucent garanties that 1Gethernet on copper vill run on GigaSpeed when
> all the standards are ready.

> Arne Keller

--
Please remove NO and SPAM from my email addr to reply.

Read what the FTC has to say about Junk Email at URL
http://www.ftc.gov/bcp/conline/pubs/online/inbox.htm
 
 
 

Gigabit ethernet

Post by john mos » Fri, 03 Jul 1998 04:00:00



>I've heard something here that 62.5/125 multimode cable cable isn't good
>for Gigabit speeds.  But then copper wasn't supposed to go that fast,
>either.  The copper problem was gotten around by using all four pairs,
>and sending and receiving on all pairs, so the actual speeds go up to
>something like 250 MHZ.  

Thanks John,

The 850nm window is not a problem at Gbs rates mainly due to the
properties of the VCSEL diodes ( narrow spectral width and qualified
launch). At the 1300 nm window there is a difference. Mind you I'm
still attempting to find reliable data.

To date it appears to use the 1300 nm devices you need a special patch
cord on the receiver side. Perhaps this is some sort of mode *!
One source has advocated the use of factory polished PC connectors
fusion spliced into the horizontal cable. This would seem to indicate
a "back reflection" issue. As you already know, this technique is not
a cost effective way to go. May as well stick with Singlemode!

For the sake of general information at the 850nm window

62.5/125 fiber 260 meters (really does better but this is where the
standard leaves it.)

50/125 fiber (for those who like to bust de facto standards)  500
meters

How much does on pay out for UTP electronics to do Gigabit on all four
pairs? Someday this will be the battleground for true cost parity!

Regards,    John Moser
Integrated Fiber Optics Inc.
(410) 647-4381

 
 
 

Gigabit ethernet

Post by Peter Guenthe » Fri, 03 Jul 1998 04:00:00



>> >>It might interest the group to know that in some initial test on
>> >>transmission for gigabit ethernet that Cat 5 was more successful than
>> >>62.5/125 micron multimode fiber.

How much did the electronics cost though??  There isn't much point in a
solution using old cable if its cheaper to pull new.

Quote:>I've heard something here that 62.5/125 multimode cable cable isn't good
>for Gigabit speeds.  But then copper wasn't supposed to go that fast,
>either.  The copper problem was gotten around by using all four pairs,
>and sending and receiving on all pairs, so the actual speeds go up to
>something like 250 MHZ.

Dead right. They couldn't get the light source to couple into the fibre
initially due to all the modes. 300 m of 500 MHz.km multimode fibre only has
a theoretical capacity of 1.6 Gbps over that length. However, much of what's
intalled to date has lower bandwidth. Getting Gigabit to work over 300m MM
OF link lengths was also apparently quite tricky.

Quote:>It seems that a design using something similar would work well on
>fiber.  Like using wavelength division, where there are two LEDs or
>LASERs to transmit, and two receivers for different wavelengths.  Sure,
>this means that the system is more expensive, but then that would be
>expected for a 10X increase in speed.  And as time goes on and the
>volume increases, the cost would go down due to mass production.

Forget about it. This is a single mode technology. You can't get bandwidth
that isn't there. Single mode is virtually only limited by the light source
itself.

Quote:>Right now, I see info that the long haul fiber circuits are going up to
>80 GBS using this wavelength muxing.  Of course they're on SM fiber, and
>the cost of a hundred thousand dollar piece of equipment is a whole lot
>cheaper than putting in miles of additional fiber.  :-)

Exactly.

Quote:>> Lucent has made a lot of testing on their new GigaSpeed cablesystem and
they
>> have 1Gethernet like signals running on it.
>> And  Lucent garanties that 1Gethernet on copper vill run on GigaSpeed
when
>> all the standards are ready.


Peter Guenther, Senior Engineer Comms
Andrew Boon Pty Ltd Consulting Engineers
PO Box 308, North Hobart TAS 7002, AUS.
Ph +61 3 6224 8277 fax +61 3 6224 8150

Web Home Page:- http://www.andrewboon.com.au

Web resources:-

Structured Cabling Guide: http://www.andrewboon.com.au/html/BOONSCS1.html
Building Comms Services Brief:
http://www.andrewboon.com.au/html/COMBRIEF.html
Fast Ethernet O'view Diagram:
http://www.andrewboon.com.au/images/100BaseT.GIF
ISP Selection Guide:  http://www.andrewboon.com.au/html/ISPguide.html
Contractor Install Guidelines:
http://www.andrewboon.com.au/html/CABLCONT.html

 
 
 

Gigabit ethernet

Post by John Lundgre » Fri, 03 Jul 1998 04:00:00


I did not write the following paragraph.

Quote:> >> >>It might interest the group to know that in some initial test on
> >> >>transmission for gigabit ethernet that Cat 5 was more successful than
> >> >>62.5/125 micron multimode fiber.
> How much did the electronics cost though??  There isn't much point in a
> solution using old cable if its cheaper to pull new.

I did write this one, tho.

Quote:> >I've heard something here that 62.5/125 multimode cable cable isn't good
> >for Gigabit speeds.  But then copper wasn't supposed to go that fast,
> >either.  The copper problem was gotten around by using all four pairs,
> >and sending and receiving on all pairs, so the actual speeds go up to
> >something like 250 MHZ.
> Dead right. They couldn't get the light source to couple into the fibre
> initially due to all the modes. 300 m of 500 MHz.km multimode fibre only has
> a theoretical capacity of 1.6 Gbps over that length. However, much of what's
> intalled to date has lower bandwidth. Getting Gigabit to work over 300m MM
> OF link lengths was also apparently quite tricky.

And I wrote this one.

Quote:> >It seems that a design using something similar would work well on
> >fiber.  Like using wavelength division, where there are two LEDs or
> >LASERs to transmit, and two receivers for different wavelengths.  Sure,
> >this means that the system is more expensive, but then that would be
> >expected for a 10X increase in speed.  And as time goes on and the
> >volume increases, the cost would go down due to mass production.
> Forget about it. This is a single mode technology. You can't get bandwidth
> that isn't there. Single mode is virtually only limited by the light source
> itself.

Well, the bandwidth over cat5 copper isn't there, either.  I guess
that's why they are using all four pairs and both transmitting and
receiving at 250 MB over each pair.

Basically the modem makers made more sophisticated modulation schemes to
get more down the phone lines: 14.4 then 28.8 and then 56.  I'm saying
that it may be possible to do something similar with FO, but at a much
higher speed, of course.

> >Right now, I see info that the long haul fiber circuits are going up to
> >80 GBS using this wavelength muxing.  Of course they're on SM fiber, and
> >the cost of a hundred thousand dollar piece of equipment is a whole lot
> >cheaper than putting in miles of additional fiber.  :-)
> Exactly.
> >> Lucent has made a lot of testing on their new GigaSpeed cablesystem and they
> >> have 1Gethernet like signals running on it.
> >> And  Lucent garanties that 1Gethernet on copper vill run on GigaSpeed when
> >> all the standards are ready.

> Peter Guenther, Senior Engineer Comms
> Web Home Page:- http://www.andrewboon.com.au

--
Please remove NO and SPAM from my email addr to reply.

Read what the FTC has to say about Junk Email at URL
http://www.ftc.gov/bcp/conline/pubs/online/inbox.htm
 
 
 

1. Fast-Ethernet/Gigabit Ethernet Diameter clarification

I know that both Fast and Gigabit Ethernet's Diameter is just over 200
Meters using UTP. However, I'm a little puzzled by what they mean by
"diameter".

I.e. I would imagine the following example has a 300 meter diameter
right?

C-----100 meters-----S-----100 meters------S------100 meters------C.

S=Switch
C=Client
(All UTP)

Would I be OK if I put MMF in the middle?

thanks,

Paul

2. showdvi save to iff

3. Dell'Oro Group: Fast Ethernet and Gigabit Ethernet Switches Drive Q1 Growth

4. HELP! Amiga at University a good idea?

5. Cisco2970 gigabit ethernet port gets disabled due to loopback error

6. Picasso II driver?

7. gigabit ethernet on Cat-6500 to Alcatel OSR problem

8. AGP440: Boot failure

9. ATM vs Gigabit Ethernet for Reliability over a WAN

10. Gigabit ethernet card for PIX 520

11. Anyone know the IEEE Designations for 40 and 160 Gigabit Ethernet?

12. 8 Port Gigabit Ethernet for Cisco 760x Routers