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

[EFM] RE: [EFM-Copper] [EFM-copper] update of carrier perspective on SH DSL presentation

Dear John and All,

I will respond on item 2 as ANSI DSM-standard editor and
someone involved in the DSM area heavily.  The ANSI DSM group
is also copied.

DSM standardization is underway with increasing contribution
from many.  There are two main categories in the document:
1. no coordination (unbundled) so no central coordination or
interaction between lines -- autonomous - today's exact situation
2. coordination (basically for remote terminals where many lines
if not all in same binder terminate on same DSLAM/box)

There is at least one large American operator very close to
deploying some early DSM methods in item 1, so that area is
not 5 years away.  Some of Doug's results are for 4 bonded
lines (2.5 Mbps each to add to 10), but also fall under category 1.
These are also near term, and at least one company has implemented
DSL systems that so operate (Voyan, San Jose) -- you might
want to see this yourself by contacting them.

There are other areas, particularly under the coordination category
that may take longer to gain acceptance.  Just as with existing
spectrum management standards -- some parts are implemented
now, some will be a while.  That is true for DSM also.

The whole
DSL field will gain from the DSM area, shortly and long term.  And,
several operators are actually putting money into this area, so its
more than just the standards people involved.
If you would like to learn more - see Chapter 11 of the book
DSL Advances - T. Starr, M. Sorbara, J. Cioffi, and P. Silverman,
Prentice-Hall: 2003.

We already had a decision on DSM in ANSI, and it passed unanimously
to go forward so all operators, vendors, chip suppliers, and a
professor/editor were in agreement.  I hope EFM will find a way
forward also that is productive for the DSL industry.  I'm pretty
confident DSM will become a strong component of that success
in the future.

John Cioffi

At 05:52 AM 1/5/2003 -0800, John.Egan@xxxxxxxxxxxx wrote:
I have some questions for you, as well. There seems to be some misunderstandings occurring here that I hope you folks from the ADSL/DMT side could clear up. These are related to the use of the "generic" term "ADSL." Is not the EFM proposal from Doug based on Annex J and not the whole family of ADSL (the generic ADSL term), as ADSL per se would not meet the Long Reach Objective (generic ADSL is low bandwidth, asymmetric, etc.)? I believe any discussion of ADSL should center not on the family of Annexes but on Annex J itself.
1) If we consider Annex J then many answers to your points immediately become apparent. For example, your item 1, which you partially answer yourself by mentioning the embedded base and investment in ATM ADSL, has a more complete answer when considering the incompatible nature of Annex J and this base. To deploy Annex J and cause service deterioration in the existing base would be foolish for an Operator.
2) As to DSM in item 2... isn't this a "house of cards" theoretical technology that would only work in a fully closed environment where only one DSL technology is deployed from a single carrier? Otherwise, wouldn't this be problematic where the service would be frequently disrupted by new disturbers added by others that are outside the DSM domain? Well, I suppose this discussion is all academic as DSM is not going to be anything more than theoretical for at least 4-5 years.
3) Your item 3 b) appears to use the generic ADSL term instead of Annex J. The incompatibility of Annex J with existing base of ADSL should be made clear. I hope that an open discussion occurs Monday as to the conflicts and incompatibilities of Annex J and not a discussion of ADSL in a generic sense, otherwise we are not doing a real comparison of Long Reach technologies.
4) By the way, what is the official title for Annex J anyway? I believe it is "ADSL for operation above ISDN." If this Annex defines service as this, does this mean we have a technology proposal based on one that would have limited in deployment in NAFTA as well (few BRIs)? To be POTS compatible versus ISDN would mean Annex J must be modified from the existing recommendation to some new DSL definition which in turn makes this a comparison of a well defined and standardized technology (g.SHDSL) and one that is not standardized (Annex J changed for POTS)?
Behrooz, Happy New Year to you. I look forward to the conclusion of this step in Vancouver and the opportunity for the Copper track to make some serious headway. I expect many at EFM would like to see the year plus delaying efforts in Copper stopped and some conclusions reached so that we finally move ahead.
-----Original Message-----
From: Behrooz Rezvani [mailto:brezvani@xxxxxxxxxx]
Sent: Sunday, January 05, 2003 1:06 AM
Cc: 'Hugh Barrass'; Howard Frazier
Subject: Re: [EFM-Copper] [EFM-copper] update of carrier perspective on SHDSL presentation

Marc thanks :-)
I have few questions for the operators (either arms of business/technical/regulatory representatives) to help me understand the situation a little bit better:
1) Do carriers think that it may not be necessary to offer Ethernet to residential market using ADSL. Could the reason be that they have invested so much into ATM/ADSL that in fact it does not make sense to change strategy at this point. In other words the ADSL volume already is so high (therefore ADSL chipset cost so low) that it would cost them more money try to introduce Ethernet over ADSL. Why mess with it when finally the inter-op is working very well and all the operational stuff with $B dollars into it is now mature
2) The other question could be regarding the performance of SHDSL:  is it as good or better than ADSL. I am not sure what is real answer to this. There are two good presentations on Monday, and to a large extend the performance data depends on the assumptions on reach, binder composition, etc. So these will go thru acid test on Monday. <However from the perspective of roadmap,  some new work and actual data based on DSM suggests for multi-pair operation the results are very good based on ADSL-dmt >
3) Is binder segregation allowed. Some operators may allow that, but in general my understanding is that will make it more expensive. This is at least the comments I received from some international operators. For example consider this: 
a) If we assume binder segregation is allowed then what is the segregation rule. Does it mean we separate all SHDSL and T1 in the same binder (symm) and all ADSLs, and POTs in other binders. In that case the ADSL binder will outperform the SHDSL/T1 binder simply because it is wider band and has more transmit power and it does not have to deal with T1 Jammer PSD.
b) If we assume that binder segregation is not allowed then it seems to me that ADSL will dominate the composition of the binder by almost 4 to 1 and in that case we have to make simulation assumptions based on those consideration, which again I would see the results favor ADSL in more cases
Anyway I don't think there are simple answers and I will be happy to receive my answers off line.
I have also made a contribution recommending to support both SHDSL and ADSL. I think this would result to a bigger footprint for Ethernet.
Best Regards
----- Original Message -----
To: ;
Cc: 'Hugh Barrass' ; Howard Frazier
Sent: Saturday, January 04, 2003 11:05 AM
Subject: [EFM-Copper] [EFM-copper] update of carrier perspective on SHDSL presentation

Since there are additional names and the file is small, we thought it would be relevant to send the latest carrier perspective on SHDSL presentation (easley_copper_1_0103) by email as well as uploading it on the server.
See you in Vancouver,

John M. Cioffi
Hitachi America Professor of Electrical Engineering
363 Packard Electrical Engineering Bldg.
350 Serra Mall
Stanford, CA 94305-9515
+1-650-723-2150  Fax: +1-650-724-3652