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

Re: [BP] NEXT and FEXT aggressor modeling




Unfortunately I was not able to attend Friday's call.  I would like to see option 1 below, your definition of deterministic.  The "special efforts" should be to simulate with the aggressor at a frequency delta to the through such that it continually sweeps the through eye and to simulate for sufficient time to gather adequate statistics, which I would say is 10M bits or more.  That's the most accurate and realistic approach to modeling crosstalk.  What I'm against is multiple simulation to analyze the same thing.  It's more modeling to be done, and the amount of channels and simulation cases we already have has become unmanageable.

Thanks,        Joe


Joe Abler                                                             abler@us.ibm.com
IBM Microelectronics Division                          919-254-0573
Technical Marketing & HSS Applications    919-254-9616 (fax)
3039 Cornwallis Road                                                                
Research Triangle Park, NC  27709



Vivek Telang <vtelang@BROADCOM.COM>
Sent by: owner-stds-802-3-blade@listserv.ieee.org

11/09/2004 04:47 PM
Please respond to Vivek Telang

       
        To:        STDS-802-3-BLADE@listserv.ieee.org
        cc:        
        Subject:        [BP] NEXT and FEXT aggressor modeling



Dear 802.3ap Task Force members (Signaling ad hoc),
 
During Friday's call, there was a discussion about the interpretation of the Straw Poll from the previous meeting which proposed that the "Next/Fext treatment should be : A. Random, B. Deterministic, C. Both<?XML:NAMESPACE PREFIX = O /> ". My interpretation of Option C (Both) was that the NEXT/FEXT aggressor would be be modeled as a deterministic source as well as a random source, in separate simulations, and that both results would be reported.
 
Mike A asked me to formalize this and post it to the reflector, in the hope that we could reach consensus on the reflector before next week's meeting. So here it is:
-------------------------------------------------------------
For the purpose of simulating NEXT and FEXT impairments and comparing their effect on various signaling schemes, the NEXT and FEXT sources shall be modeled in the following two ways, and both results shall be reported:
1. Deterministic: A random data signal of the type of 10GBASE-KR
2. Random: Gaussian noise filtered to have the same PSD and power as 10GBASE-KR
--------------------------------------------------------------
 
Rationale:
Both models have unique advantages and disadvantages from the point of view of evaluating the competing signaling schemes. The deterministic source is most like a real-life aggressor, but the results will be somewhat anecdotal, and special efforts will need to be made to make it behave in a worst-case manner (aligning the phase with the received signal, etc.), especially when modeling multiple aggressors. The random source does not look like a data signal, but rather represents a collection of asynchronous sources. It is simpler to model, which is important in our context, since we want to minimize the number of parameters that can be interpreted differently.
 
Comments?
 
Regards,
 
Vivek
 
 
Vivek Telang
Broadcom Corp.
(512) 651-9523

vtelang@broadcom.com