Re: [802.3_EPOC] Channel Model meeting minutes
Marek,
I think a link model tool would be desirable if one that properly models the coax plant can be defined. I believe the table approach is complementary to that. The tables I sent to Duane were only meant to start a list of some RF characteristics that clearly impact OFDM/OFDMA and that we should be asking the MSO community for input on (i.e. levels). I am sure there are others but we must be careful not to overspecify by adding RF characteristics that are overlapping or redundant. For example, MER, SNR and NPR all are influenced by the AWGN in a channel but have different nuances.
In time, a link model tool might even eventually supplant the table but note that MSOs are used to defining plant conditions via tables like the ones in the minutes. DOCSIS was built with the channel model defined in table form. In addition, only OFDMA system that was actually deployed in volume on HFC (with over 1 million modems fielded) was developed using a table very similar to the simple one I sent.
I worry that the HFC system is much more complex than the optical system that was modeled by the Excel tool for EPON. A comprehensive channel model tool that takes into account various noise sources, linear and non-linear impairments and accounts for the fact that there are other signals (analog and digital video, DOCSIS etc) already riding on the coax may take a long time to specify and build.
I searched briefly to see if I could find a tool that was defined for the development of LTE standards for some guidance on this. I did find a number of tools from vendors or specific tool for simulating MIMO, but nothing that came out of 3GPP or ETSI directly (someone please chime in if there is such a thing). However even LTE, which had a much larger technical community contributing to it, still used channel models in table format such as this reference from ETSI TS 125.104 http://www.etsi.org/deliver/etsi_ts/125100_125199/125104/08.09.00_60/ts_125104v080900p.pdf
[cid:image002.jpg@01CDA87B.47E9A270]
We have to walk before we can run, and I think a table has to be our starting point.
Hal
From: Marek Hajduczenia [mailto:marek.hajduczenia@xxxxxx]
Sent: Friday, October 12, 2012 10:29 AM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] Channel Model meeting minutes
Duane,
Sorry for beating on you again about this, but are you sure that the list of attendees is complete? When I rejoined the call after it dropped, I was informed there were 26 people on the call. I do not count 26 people in the list.
It is not clear to me why people join the calls and have problems disclosing their affiliation. Is this a big secret or something? Even if someone is joining the call just to follow-up on what's going on, the rules are the same for everybody - let the Chair know who you are and who you are affiliated with.
As for the minutes, you are missing the important (I think) discussion on the whole purpose of the ad-hoc at large. I believe (as indicated on the call) that tables are nice to have, but they do not provide the link model tool we are used to 802.3. I recall that many of the problems related with PHY design in EPON were solved much easier because people had at their disposal a simple tool in Excel, where they could plug in the numbers and see whether their assumptions would work with a very good level of confidence. It did not only simplify the discussions (everybody was using the same model) but also facilitated proposals, where people would bring whole tables of parameters for Tx and Rx devices derived from the approved model. I fear that if the link model is limited to a few non-interactive tables in Word or even Excel, with no facilities provided to modify them and see results of such changes, the task of this ad-hoc will be largely missed. I would urge the group to consider building a simple (and potentially approximated, due to limited computational capabilities of Excel) tool allowing people to experiment with the link definitions, observing impact of different configuration on link feasibility, BER levels, SNR etc.
If needed, I can provide the necessary Visual Basic expertise to write the required macros and front-end.
Note that this does not mean that we should forget about using Matlab(-like) tools for precise simulations. However, I believe it is possible to generalize them and create approximate models in Excel as well, for the general consumption by the task force.
I am also concerned about the potentially proprietary nature of Matlab models, as mentioned on the call. To me that means that results coming from group A may be potentially incompatible with results coming from group B. We need to be comparing apples to apples when taking decisions. For that purpose, I would urge people to consider sharing the Matlab model (if used) so that it can be scrutinized by the Task Force and potentially approved as an official link model tool.
Regards
Marek
From: Duane Remein [mailto:Duane.Remein@xxxxxxxxxx]<mailto:[mailto:Duane.Remein@xxxxxxxxxx]>
Sent: Friday, October 12, 2012 01:12
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx<mailto:STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx>
Subject: [802.3_EPOC] Channel Model meeting minutes
All,
Please find attached the unapproved meeting minutes for today's call.
As always, please let me know of any corrections.
Mark,
Please post these in the Ad Hoc area.
Best Regards,
Duane
FutureWei Technologies Inc.
duane.remein@xxxxxxxxxx<mailto:duane.remein@xxxxxxxxxx>
Director, Access R&D
919 418 4741
Raleigh, NC
________________________________
________________________________
<="" p="">
________________________________________________________________________
To unsubscribe from the STDS-802-3-EPOC list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-EPOC&A=1
