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

Re: [EFM] RE: [EFM-Copper] Copper schedule in LAX (andbeyond...)


    getting an agreement on noise models and loop characteristics for
performance evaluation doesn't necessarily mean we need to vote for the line
code. As a matter of fact, we got these agreements in Europe and US and still
don't plan to pick the line code right now. Also, different line codes we keep
in mind give almost the same performance just because they all are QAM. The
difference is mostly in duplexing and implementation issues. The latter maybe
requires some more field experience (despite Etherloop, 10Base-S, ADSL, TDD VDSL
and SCM VDSL already has a good one). The first, I think, might be selected
rather soon. We got requirements and need to see how to accomplish them. We need
a simulation environment agreed.

      Another possibility is to leave physical layer to Q1/03 and move to
TPS-TC, Layer 2/Layer 3. Fine with me as well. The same thing is the 4 ADSL case
- this idea also doesn't require to specify anything for physical layer. We use
what is done and go straight to Layer 2 and 3, as Mark Kimpe is proposing. It
will be not easy both to find 4 pairs to the same residence and to pay $20 per
pair before the service starts, but if service providers like that, I like that
The small point is, I have got a concern that 4 ADSL lines will ever give 10
Mb/s symmetric for 12 kft; seems even 5 Mb/s is very difficult to accomplish
using "conventional" means. But this is really small.


Behrooz Rezvani wrote:

> Vladimir, Patrick, all,
> First of all, I would prefer a decision for line code the sooner the better.
> Here are some considerations though:
> I think if we are dealing with Public network only, then it is much easier
> to deal with issues you brought up. After all for short public loops (i.e.
> 2.5 kft) VDSL and 100Base-CU both are saying that they will comply with
> plans 998/997. As everybody in this committee may know that there is a dual
> track std in ANSI for VDSL due for a decision in Q1/03. Mean time TDD
> methods can meet the spectral compatibility and qualify to be used in public
> network. The reason it was decided that we go through this process and make
> a call by Q1/03 was because of the field experience, and to my knowledge
> nobody has rollout VDSL 998/997 yet for field trial purposes. In order to be
> consistent with the roots of VDSL, 998/997/TDD I suggest to plan for line
> code vote in similar time frame as ANSI. Vlad, I don't think that equipment
> vendors are going to deliver in middle of 04, rather, I think people will
> start field trial by Q1/Q2-02 based on these technologies.
> The only new method proposed here that is very valuable and different is 10
> Base-T4 which delivers 10 Mbps over 12 kft (AWG-24) and has in fact traction
> with RBOCS. Because there is a real business case behind it. I think if we
> like to pursue that as a group I will certainly support to move faster, i.e.
> Q4/02 for a decision, whatever Howard, Hugh, and the rest of the team can
> manage to mobilize. But there we have pay a lot more attention to Spectral
> Compatibility because of a lot basis services are co-existing with this
> service in the bundle.( I don't think that would be an issue)
> Now, if we also like to focus on in building wiring, I think there is little
> data, measurement, or simulations of channels that we have access to, or
> they are unavailable. High frequency FEXT and NEXT still is a problem,
> although Howard in our last conference call pointed us to 10 Base-T and in
> bundle characterization.
> Thanks
> Behrooz
> -----Original Message-----
> From: Stanley, Patrick [mailto:pstanley@xxxxxxxxxxx]
> Sent: Tuesday, October 16, 2001 8:09 AM
> To: 'Vladimir Oksman'; Behrooz Rezvani
> Cc: 'Hugh Barrass'; 'Howard Frazier';;
> Subject: RE: [EFM-Copper] Copper schedule in LAX (and beyond...)
> Vladimir,
> I agree. You bring up a good point. As I state in my presentation that I
> will make this week, the NEXT and FEXT models currently used in the
> standards are based on a large number of measurements (>90,000). Reproducing
> this level of measurement in order to replace the models is a substantial
> amount of work, and could slow down development of the standard.
> Regards,
> Patrick
> -----Original Message-----
> From: Vladimir Oksman [mailto:oksman@xxxxxxxxxxxx]
> Sent: Tuesday, October 16, 2001 10:32 AM
> To: Behrooz Rezvani
> Cc: 'Hugh Barrass'; 'Howard Frazier';;
> Subject: Re: [EFM-Copper] Copper schedule in LAX (and beyond...)
> Behrooz,
>      I think it is definitely a good idea to make a complete investigation
> for
> the loop plants, crosstalk coupling models and noise sources all over the
> world.
> However, it seems to me that we couldn't complete this huge work not even to
> Nov
> 02, but even to Nov 03 on one hand, and, on the other hand, even Nov 02
> sounds
> too late. It actually targets us for a decisions regarding duplexing and
> modulation technique to the middle of 2003 and for equipment, accordingly,
> to
> the middle 2004! I am not sure many people are looking for this long
> process.
>       My view is, we can start right now from public networks, which are
> rather
> well known and specified, add all the available data on loops inside big
> buildings and take this package as a primary environment to target EFM.
> Later
> on, as different new environments will show a good business case, we just
> add
> them to the existing models and improve the spec.
> Vladimir
> Behrooz Rezvani wrote:
> > Hugh and Howard, All,
> >
> > Sorry that I have to leave Thurs afternoon for Stanford Conference on
> > Copper, but here are some thoughts:
> >
> > one of the areas that I am interested to focus on is first and for most
> the
> > simulation environment. We have had couple of conference call on this
> > subject in the past two months and we have made good progress.
> >
> > Here is the summary of what is needed (from everybody) and I certainly
> like
> > to see good contributions as I am trying to put EFM data rate analysis
> > document together:
> >
> > We agreed we need to break copper loops in to two areas:
> > Public Plant/Network and MxU Network
> > We are collecting input right now and two groups are working together to
> put
> > this in to play:
> >
> > The first Item that we like to address is public loop plants. We hope to
> > have an early draft document in Austin to discuss this. Hopefully by
> January
> > interim meeting we can close this! OK may be.
> >
> > Next we like to have an early draft of MxU loop plants in the presence or
> > absence of public loops. This is a new area to everybody. It include
> issues
> > such as cable models, interaction of different cable types, etc. Based the
> > early work that we are reviewing we like to close this by March or May
> time
> > frame (:))
> >
> > Next topic is modeling of all noise sources interferences and inclusion of
> > all noise sources. A lot of Next and Fext models don't work very well.
> This
> > is important because we need more upstream bandwidth. Spectral
> compatibility
> > and Dynamic spectral Compatibility, stationary noise and non-stationary
> > noise models are mostly new work that needs to be taken in to account. I
> > really wish that we might get a handle of that and get some sense of it by
> > May 02, may be even Nov 02.
> >
> > This is the ground work that we need to do allow simulation to be
> evaluated
> > fairly. The plan is to use as much work that is done by ANSI, ETSI and
> > ITU-T. I have been in touch with various group in these organization to
> get
> > their participation here at EFM and we could all benefit from that.
> >
> > Thanks
> > Behrooz
> >
> > -----Original Message-----
> > From: Hugh Barrass [mailto:hbarrass@xxxxxxxxx]
> > Sent: Friday, October 12, 2001 11:12 AM
> > To:
> > Cc:
> > Subject: [EFM-Copper] Copper schedule in LAX (and beyond...)
> >
> > All,
> >
> > First of all, I would like to apologize to all those who received two
> copies
> > of this
> > e-mail. If you are interested in the copper track and didn't receive two
> > copies then
> > you need to register for the copper mailing list.
> >
> > If you are not interested in the copper track then stop reading here...
> >
> > Follow this link to read instructions for the mailing lists.
> >
> >
> >
> > Finally, the payload data....
> >
> > In LAX we will have 1.5 days of meetings to ourselves. The EPON people
> will
> > leave us
> > alone and we can concentrate on Resistance, Inductance, Current Density
> and
> > other
> > concepts which would otherwise upset the optical folk. Currently we have
> 15
> > presentations submitted covering rate and reach; theoretical analysis; PMD
> > proposals;
> > general requirements and 1 new objective. My estimate is that we will
> > comfortably
> > cover these presentations on Wednesday afternoon and Thursday morning -
> > leaving
> > Thursday afternoon for free-range discussions before we rejoin the group
> on
> > Friday
> > morning.
> >
> > I would like to use this "spare time" to (pre) discuss the new objective
> and
> > the
> > simulation / test ad hoc. I would also like to allow some time for a
> > breakout to
> > discuss rate mechanism / operation (as I mention in my recap presentation)
> > and any
> > other (copper) business.
> >
> > Time is very short, so it will be hard to prepare for the Austin Plenary.
> I
> > suggest
> > that we use the spare time and gaps to work towards the baseline
> > presentations that
> > will be needed in November. This means that you should all be trying to
> seek
> > out like
> > minded people and sharing your ideas for November presentations.
> >
> > I will send a copy of my recap presentation on the copper reflector - it
> > will also go
> > onto the website.
> >
> > Hugh.