Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: 850 nm solutions


Your note seems to imply that Hari was developed within Infiniband and then
introduced from
there into 802.3. This is not my understanding of its history. The
Infiniband group developed/
is developing a 2.5 Gbaud/s serial link for use in 1-wide, 4-wide, and
12-wide configurations
using the 8B/10B code. Somewhat in parallel with this, people from the Fibre
Channel and 
Ethernet communitties got together to look at what might be good interfaces
to use between
physical layer chips for 10 Gbit/s implementations and came up with Hari and
Sali which
are roughly equivalent to the current proposals for XAUI and XGMII. These
people also
chose the 8B/10B code for Hari. Since one 4x2.5 Gbit'isn 8B/10B interface is
pretty much
like another, there is similarity between Hari and the Infiniband x4
interface though
there is a 25% speed difference. 

The interfaces were each developed by communities focused on their market's
needs. In my
opinion, the decision to use different speeds was driven by differences in
the respective
market needs. 

An interface at these speeds is analog. This is particularly true if it is
to serve the
length of traces likely to be found between transceivers and switch chips.
Taking analog
considerations into account when we develop the standard will enable
robust designs. XAUI is very suitable to the use for which it has been

The point of my note was: if we were going to standardize a short run copper
link, it
would make sense to look at what could be done on a 4-wide connection vs. a
10 Gbit
serial connection. Our existing decision has been to not do a short copper
link -
probably driven in part by the low usage of 1000BASE-CX. 

Pat Thaler 

-----Original Message-----
From: Roy Bynum [mailto:rabynum@xxxxxxxxxxxxxx]
Sent: Wednesday, April 26, 2000 6:43 PM
To: THALER,PAT (A-Roseville,ex1); Rick Walker; stds-802-3-hssg@xxxxxxxx
Subject: Re: 850 nm solutions 


For Infiniband,  I think that HARI is a very good solution.  I question the
way that it was introduced and developed as part of the
effort in something that is not Infiniband.  If people want to make products
for Infiniband, I have no problem with that.  As a
customer, I question the motivations of my vendors to have me pay for the
development of technology that was actually intended for
another use.  I wonder how much that has already increased the price of the
product that I will be receiving.  I wonder even more
how much the vendor was actually trying to develop something for my use
instead of somebody else, and gave me, the customer, the
"left overs".  I wonder how much better the product, that I may buy, would
have been better if the vendor had not been developing
technology for another use.

As a customer, I was hoping to receive an 802.3 Ethernet product that
treated the interface to the optical domain as a digital
optical system, not an analog copper system, which you refer to for the use
of HARI.  As a customer I was hoping that the vendors
would listen to me and my requirements and look at it as an opportunity to
enter a market that is as large as the global Internet,
instead of staying in the collective enterprise space.  Vendors that are not
looking at the market correctly have already lost their
market share in the Internet backbone, and they are about to start loosing
it at the access edge as well.  History has shown that
customers will get what they want one way or another.

The response of a BIG customer,
Thank you,
Roy Bynum

----- Original Message -----
From: THALER,PAT (A-Roseville,ex1) <pat_thaler@xxxxxxxxxxx>
To: Rick Walker <walker@xxxxxxxxxxxxxxxxx>; <stds-802-3-hssg@xxxxxxxx>
Sent: Wednesday, April 26, 2000 1:08 PM
Subject: RE: 850 nm solutions

> Infiniband will be using something very similar to the HARI interface over
> short copper links though the distance goal is, I think, 6 m. To travel
> short copper cables, it may make sense to use a 4 wide signal from HARI
> rather than 10 Gbit/s serial.
> -----Original Message-----
> From: Rick Walker [mailto:walker@xxxxxxxxxxxxxxxxx]
> Sent: Wednesday, April 19, 2000 4:58 PM
> To: stds-802-3-hssg@xxxxxxxx
> Subject: Re: 850 nm solutions
> > Jim Tatum writes:
> > But why does it matter? Why limit the users? Why not put in the table.
> > costs nothing.  Just put in what the model and data tell us to.  It is
> > my opinion that a large percentage of 10GB style links are going to be
> > very short, less than 10m.  If you look at the way many fiber ports
> > are being used today, many are in the 10m range.  Also, since copper
> > cables are going to be EXTREMELY challanged to go that distance at
> > 10GB, why not let the market choose the lowest cost solution using
> > 850nm VCSELs and 62.5um fiber?
> FWIW, I agree that 10G across CAT-6 or other twisted pair would be very
> difficult.  However 10G across coaxial cable is fairly easy.  It can be
> done with 0.1" diameter coaxial cable using simple NRZ data encoding.  A
> simple FIR pre-equalizer can double this distance.  Without a doubt
> copper would be the cheapest solution for links under 10M.  I would
> estimate a mature chipset price of about $50 per end and $15 for the
> cable.
> This performance was demonstrated in 1998 using a 25GHz bipolar chipset.
> See: Walker, R. C., K. Hsieh, T. A. Knotts and C. Yen, "A 10Gb/s
> Si-Bipolar TX/RX Chipset for Computer Data Transmission" , ISSCC Digest
> of Technical Papers 41(February 1998), 302,303,450.
> A Copper PHY was voted down by the committee because it was thought that
> there was no market for this type of low-cost short distance link.
> kind regards,
> --
> Rick Walker