[opendtv] Re: Spectrum is too valuable

  • From: Craig Birkmaier <brewmastercraig@xxxxxxxxxx>
  • To: opendtv@xxxxxxxxxxxxx
  • Date: Tue, 17 Nov 2015 10:30:27 -0500

On Nov 16, 2015, at 9:56 PM, Manfredi, Albert E
<albert.e.manfredi@xxxxxxxxxx> wrote:

How do you see these as being the same link?

You are correct. The links are different. But the information about DOCSIS was
the same. No doubt this is why I did not see the information about the long
term goal for the standard is to increase the bandwidth of the cabled systems
to 1.788 GHz.

But that is irrelevant, as I am not aware of any system that operates at 1.788
GHz today, or even the 1.218 GHz, which "may" be possible with existing HFC
systems by upgrading the amps in the PONs.

But most important, the higher band pass and theoretical limits of DCOSIS 3.1
are irrelevant to your original point, which was that the MPEG-2TS channels
assigned to digital cable should be recovered now for broadband.

Obviously, you did not look at the standard. That's why you did not know that
DOCSIS 3.1 depends on having a whole lot more cable banswidth than DOCSIS 3.0
could use.

That is entirely false. What's more, you could significantly increase the total
downstream bits available using DOCSIS 3.0 in a 1.788 GHz HFC system.

As both Ron and I pointed out, along with other papers I have linked to, DOCSIS
3.1 will be deployed next year alongside analog cable, digital cable and DOCSIS
3.0 services in 1 GHz HFC systems. All that is required is a band of spectrum
that occupies ANY portion of the HFC spectrum above 54 MHz; that band can be a
minimum of 24 MHz and a maximum of 192 MHz.

If you doubt that, look at your PHY standard. But be forewarned, the standard
is self contradictory, or at least offers exceptions, with respect to what is
"required" for implementation. For example:
7.2.1 Downstream CM Spectrum8

The CM MUST support a minimum of two independently configurable OFDM channels
each occupying a spectrum of up to 192 MHz in the downstream.

The demodulator in the CM MUST support receiving downstream transmissions up to
at least 1.218 GHz. The demodulator in the CM MAY support receiving downstream
transmissions up to at least 1.794 GHz.

The demodulator in the CM MUST support agile placement of the OFDM channels
within the entire supported downstream range.

Operators may need individual CM implementations to limit the spectrum over
which the CM is able to receive downstream signals in order to coexist with
other network signals above 1 GHz.

As a result of this operational constraint, the CM MUST support one or more of
the following downstream upper band edges: 1.002 GHz, 1.218 GHz, 1.794 GHz.
Support in this context includes the capability to demodulate up to the
supported band edge. This operational consideration does not modify the
requirement for the demodulator capabilities; it only acknowledges that certain
configurations may need to operate at a slightly lower upper band edge due to
MSO-specific operational constraints.

Not until Ron supplied you with some viewgraphs, that showed you what you
should have read previously.

My calculations were entirely consistent with the view graphs Ron supplied.
Unlike myself, he had the good sense to distance himself from an argument with
you, and simply provided some very useful information debunking your claims
about the standard.

By the way, did you notice in your first article today how Cox is NOT
changing out its cabling for DOCSIS 3.1?

Wow! Imagine that!

They are doing just as I claimed, adding support for DOCSIS 3.1 within their
existing 1 GHz plants. Here in Gainesville it will look very much like one of
the slides Ron provided: HFC spectrum Minimal available for U.S.

Just wanted to make sure you weren't going to respond that the article said
no such thing. Your assertion that all the cabling would need to be changed
was a pure guess, Craig, and apparently not true.

The article said no such thing. It said what I have been saying all along, that
Cox will deploy DOCSIS 3.1 in their existing 1GHz HFC systems.

You are the one who pointed out that the standard "supports" both 1.218 and
1.788 GHz HFC systems, and that to reach the theoretical limit of 10 Gbps DS,
the entire spectrum (above the US band) would need to be dedicated to DOCSIS
3.1 channels.

But the discussion was NEVER about reaching the theoretical limits of DOCSIS
3.1. It was about recovering the spectrum now used for analog and digital tiers
to increase the broadband throughput. I simply did the calculations for what is
possible with DOCSIS 3.1 in today's 1 GHz HFC plants. AND I was generous,
giving the best case number: 7.8589 bps/Hz

When you came back with the theoretical limits from the standard, I introduced
the information about the issues in upgrading to a 1.788 plant. I also noted
that SNR within the area served by a PON would be a limiting factor.

This is in fact one of the advantages of DOCSIS 3.1, in that it allows the
tuning of the downstream bands to the actual conditions in the PON, using less
sensitive constellations where the SNR is not adequate to support the
theoretical limits. Another advantage is that within a DOCSIS 3.1 band, some
portions can still use the QAM modulation needed to support legacy DOCSIS 3.0
modems.

It is not a guess that cabling may need to be replaced to reach the full
theoretical limits of DOCSIS 3.1 in a 1.788 GHz plant. RF emissions are a major
problem for cable systems - doubling the frequencies used is going to result in
at least some of the existing cabling requiring an upgrade.

Cox just replaced the drop from the street to my house, upgrading the coax in
the process. That alone provided a very large boost in my broadband speeds, not
to mention a huge improvement in the SNR for the analog TV service; and that
was just to realize the capabilities of the 1 GHz plant.

The industry had to upgrade the entire physical plant to move from the legacy
~400 MHz plants to the 900 MHz and 1 GHz plants deployed in the late '90s, at a
cost of several hundred billion dollars.

I have no idea what the 1.788 GHz upgrades will cost; if new cabling is
required throughout the neighborhood served by a PON, they may just deeply
fiber to the home, as they are doing in new neighborhoods. Thus it's quite
possible that the higher DOCSIS 3.1 rates may simply be leap-frogged by fiber.

Yes, Craig. See above. You deploy a lot of extra broadband capacity, without
having to install new cabling and changing the topology of the PON. Am I
beloboring the obvious or what?

You are continuing to change the argument.

Clearly DOCSIS 3.1 is compatible with existing 1 GHz HFC plants. These plants
"might" even be able to support the 1.218 GHz upgrade. It is far from clear if
the 1.788 GHz upgrade will be possible without many other labor and capital
intensive upgrades.

I have had multiple choices since the late '90s Bert.

Ooops. I spoke too soon. Now Craig is trying to convince us that he was
connected to three MVPDs.

No Bert. I am stating the obvious. There has been competition in MVPD services
since DirecTV launched in 1995, and the telcos started overbuilding cable
systems shortly thereafter.

It is only the mandated-to-be-neutral Internet service that is giving you
virtually unlimited choices of counte source, Craig. That Internet you didn't
want to be neutral.

I have had virtually unlimited choice in content sources since the '80s, first
with a VCR, then later via DVD. The Internet has not significantly expanded
choices, unless you like to watch cute cat videos. It has made it possible to
easily view content from other countries, and music videos. And it has rendered
earlier technologies like the VCR and DVD to the scrap heap of history. Now it
is rendering the glut of rerun channels in the MVPD bundles irrelevant as well.

You were not watching entire programs.

I was watching entire programs, with very, very few ad breaks back then, and
each ad break only 30 seconds long. You have no clue, Craig. Get your facts
before attempting to argue.

You are mistaken.


Regards
Craig

Other related posts:

  1. Home
  2. opendtv
  3. Archive
  4. 11-2015