RE: Short haul PMDs
Based on my own experience, there has been no real cross effect in recent
history between 1000BASE-T and -CX. During the product planning I was
involved in from 1997 to early 1999, decisions were made about CX over
Cross effect, if any has happened more recently when design teams and
supply chain managers could actually get samples and competitive prices
from the long list of 1000BASE-T PHY vendors. The small amount of CX parts
used for gbE external backplane interconnect of stackable switches might
well disappear, as 1000BASE-T PHYs drop in price and 1000BASE-T gets
designed into the next generation of external backplanes for stackable
i dooubt that considerations of -CX are not going to help us define
objectives for very short links of 10 GbE in the equipment room, except
maybe to point out that 802.3 has been wrong once already when trying to
define this application space and specify media types and connectors that
only make customers' lives more complex and do not gain any traction in the
market place. I still remember explaining to customers that CX used a
cable that was not coax, was not twinax, was not IBM Type 1....
At 12:20 PM 8/7/00 -0700, Geoff Thompson wrote:
>I would add to your point that much of the anecdotal conversation that I
>have actaullay heard over the years since "z" has been that to NOT use CX
>for short haul but rather use 1000BASE-T in that application space on the
>assumption that 1000BASE-T was going to be a "volume" product where CX
>presumably was not. I always assumed that folks were believing that
>1000BASE-T would make significant volume in the 100m in-the-wall
>application space. That is an assertion that has not been realized to date.
>The cross effect of CX on 1000BASE-T and vice versa is something that
>should be included in the conversation when/if we reopen the objectives
>for short-haul and talk about broad market potential and economic feasibility.
>At 10:33 AM 8/7/00 -0600, firstname.lastname@example.org wrote:
>>Perhaps I did not make myself clear enough. Much of the short haul
>>discussion has been about links much shorter than 100 m. 1000BASE-T
>>reaches 100 m. The question was what 1000BASE-CX numbers look like.
>>From: Marshall Eisenberg [mailto:email@example.com]
>>Sent: Friday, August 04, 2000 6:45 PM
>>To: Seto, Koichiro; firstname.lastname@example.org; email@example.com;
>>Subject: RE: Short haul PMDs
>>According to Dell'Oro's numbers, about 8,000 1000Base-T networking ports
>>shipped in Q1/00. The Q2/00 numbers should be out soon.
>>[mailto:firstname.lastname@example.org]On Behalf Of Seto, Koichiro
>>Sent: Thursday, August 03, 2000 4:58 PM
>>To: email@example.com; firstname.lastname@example.org; email@example.com
>>Subject: Re: Short haul PMDs
>>[Date: 08/03/2000 From Seto]
>>Maybe, HP pro-curve unit has... They are the one of a few vendors actually
>>shipping 1000BASE-CX options I believe.
>> > So, does anyone have data on what percentage of Gigabit Ethernet links
>> > ship with 1000BASE-CX?
>> > Pat
>> > -----Original Message-----
>> > From: Geoff Thompson [mailto:firstname.lastname@example.org]
>> > Sent: Wednesday, August 02, 2000 4:45 PM
>> > To: Hakimi, Sharam (Sharam)
>> > Cc: stds-802-3-hssg
>> > Subject: RE: Short haul PMDs
>> > Sharam-
>> > I think 15 years is a slight exaggeration. 10BASE-T started as a Study
>> > Group in August of 1987. It was the first group to have a 100 meter
>> > objective.
>> > Before 10BASE-T there was the following:
>> > 10BASE5 + AUI The AUI cable was 50 meters max
>> > 10BASE2 The coax was daisy-chained w/ an overall max of 185m
>> > 1BASE5 (StarLAN) The hub to DTE distance was 250 m
>> > 10BROAD36 Broadband diameter of 2800 meters
>> > FOIRL 1000 meter links
>> > The point here is that topology has evolved as well as speed.
>> > We have in the recent past had a distance objective for short cables in
>> > 802.3z (ref: 1000BASE-CX).
>> > The 100 meter distance history is rooted in facilities cabling and is
>> > on the (literally) cast-in-concrete distance of 90 m max from the cabling
>> > closet to the telecommunications outlet.
>> > If the prime market for an Ethernet project is not oriented to
>> > "in-the-wall" cabling then the 100 meter distance is not sacrosanct. The
>> > objective for 100 meters as approved in York was for "installed" cabling
>> > from ISO/IEC 11801. That means that we were talking about generic
>> > facilities cabling at the time, not application specific cabling.
>> > If we do go to application specific cabling then we have to do the
>> > specification in our own standard instead of referencing an outside
>> > standard such as ISO/IEC 11801.
>> > Geoff
>> > At 11:16 AM 8/2/00 -0400, Hakimi, Sharam (Sharam) wrote:
>> > >During the past 15 years and through all of 802.3 distance Objectives,
>> > >meters has been the minimum and essentially the trademark of IEEE 802.3.
>> > >There has always been discussions that if the distance is reduced we can
>> > >provide less expensive PHYs, but the cost difference never justified
>> > >development of these PHYs . Providing 100 meter solutions does not
>> > >anyone from using a 10, 20, 30 or other length cables as their needs
>> > >require. However, if time has come that the cost difference between a 100
>> > >meter solution and something less will justify such development then we
>> > >could look at it later, but changing the objectives at this time is a BAD
>> > >idea in my opinion.
>> > >
>> > >Sharam Hakimi