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

RE: 1000BASE-T PCS question


Richard has a good point which I agree. 

In fact, as I mentioned before, in order to maintain the advantage of data
rate increase by 10 fold --  from GbE to 10GbE -- the BER should be better
than GbE's BER of 10^-12.  Otherwise the retry-time will negate the
advantage gained from 10GbE.  I may try to present the analysis in July

To furthermore enhance Richards's error table, every time system get an
error, the retry-routine will takeover which is very time consuming and
wasting; as a result, the penalty is no just frequent errors, but also very
noticeable delay added by re-try.  

Ed Chang
Unisys Corporation 

-----Original Message-----
From: Rich Taborek []
Sent: Thursday, May 27, 1999 2:06 PM
To: Jaime Kardontchik
Cc: HSSG_reflector
Subject: Re: 1000BASE-T PCS question


A point of clarification: Ethernet standards, at least the IEEE 802.3
ones, specify several BER objectives, usually depending on the specific
PHY. 1000BASE-X specifies a BER objective of 10 E-12. This PHY is based
on the Fibre Channel PHY which specifies an end-to-end system BER of 10
E-12. In practice, many of the optical links I've worked on over the
years exhibit BER well below this figure. 10 E-12 is generally a
cost/testability tradeoff floor where getting a transceiver vendor to
sell you a 10 E-13 instead of 10 E-12 transceiver may result in your
forking over a considerable additional chunk of change.

I'm assuming that testing transceivers at 10 E-13 may be a bit easier at
10 Gbps than at 1 Gbps. If this is true, this means that may be able to
set BER objectives for the 10 GbE more aggressively than for 1 GbE,
possibly closer to 10 E-13.

To put BER into perspective:
A system BER of 10 E-10 @ 10 Gbps = a bit error every 1 second.
A system BER of 10 E-12 @ 10 Gbps = a bit error every 1 minutes, 40
A system BER of 10 E-13 @ 10 Gbps = a bit error every 16 minutes, 40


Jaime Kardontchik wrote:
> Ed,
> I think that other people might be better qualified than I to handle this
> question.
> My impression is that it would have been much more simpler if the group
> been called "10-Gigabit-Ethernet" instead of "Higher-Speed Group". At some
> time in the future the group will have to decide what kind of standard to
> on.
> Ethernet standards deliver a BER of 10^(-10).
> Jaime
> Jaime E. Kardontchik
> Micro Linear
> San Jose, CA 95131
> email:


Best Regards,

Richard Taborek Sr.    Tel: 650 210 8800 x101 or 408 370 9233       
Principal Architect         Fax: 650 940 1898 or 408 374 3645
Transcendata, Inc.           Email: 
1029 Corporation Way     
Palo Alto, CA 94303-4305    Alt email: