& cplSiteName &

The Ultimate Cable Modem

Jeff Baumgartner
10/27/2009
50%
50%

5:40 PM -- DENVER -- CTAM Summit -- The CableLabs Docsis 3.0 specs call for a modem to be capable of bonding, at a minimum, four upstream channels and four downstream channels. There's not much actual upstream channel bonding occurring in the wild yet, but that 4x4 configuration ensures that MSOs can offer burst speeds of more than 100 Mbit/s in both directions.

Not lame. But not where things are heading.

These days, vendors are working on modem configurations that bond eight or 16 downstream channels, while maintaining a four-channel upstream, Carey Ritchey, general manager of Microtune Inc. (Nasdaq: TUNE)'s cable unit, mentioned to me yesterday. At full burst, a 16-channel downstream (using 6MHz-wide channels) would generate something like 640 Mbit/s.

Definitely not lame. But also not where some see this all going... at least among the mad scientists out there.

Ritchey said he's already seen some requests for Docsis modems that can bond 32 downstream channels. My cable math suggests that would put a burst speed in the neighborhood of 1.2 Gbit/s.

"But that's been the extreme," Ritchey says. It's the type of extreme Cisco Systems Inc. (Nasdaq: CSCO) openly conceptualized in January 2008 at the Society of Cable Telecommunications Engineers (SCTE) Conference on Emerging Technologies with a presentation titled "A proposal for Docsis 4.0."

Seems as though that vision has since taken a few steps closer to reality.

Of course, the question is what sort of extreme would require such a massively wide speed pipe. If I had to guess, it would have to be for something along the lines of a QAM-IP gateway that can do everything today's digital set-tops can now, but also help MSOs pursue a unicast video model in a big way. Someday.

Or maybe it's about something completely different. But what? I'll do more asking around here. Got an idea? Please share it on the message board.

— Jeff Baumgartner, Site Editor, Cable Digital News

(26)  | 
Comment  | 
Print  | 
Newest First  |  Oldest First  |  Threaded View        ADD A COMMENT
<<   <   Page 2 / 3   >   >>
Stevery
50%
50%
Stevery,
User Rank: Light Beer
12/5/2012 | 3:53:19 PM
re: The Ultimate Cable Modem


@ Seven:&nbsp; for certain kinds of noise.


I should have stated that my schoolboy analysis is premised on AWGN, which I grant you has been observed exactly once (in the hallowed halls of Murray Hill during 1968 iirc).&nbsp; For that case, I believe Shannon's result says you cannot improve the capacity using any form of FEC.




@optical:&nbsp; So let's be conservative and say you have only 1ghz of useful coaxial bandwdith (instead of 3), and let's say you can only push 64 qam (6 bits/s/hz) over the spectral range, that gives you raw bandwidth of 6gbps.


So the meat of my question is:&nbsp; How close to the limit is your scheme?&nbsp; What prevents me from taking the 64 QAM up to 65,536 QAM?&nbsp; And in answering that question, I suspect the answer involves 2 numbers:

<ol>
<li>The bandwidth</li>
<li>The SNR</li>
</ol>

Now, seven's comment might be pointing at a different class of answers.&nbsp; That class goes something like:&nbsp; Stevery is asking the wrong question, because the system cannot be modeled as simply as Shannon did.&nbsp; (Analogy in the optical world:&nbsp; Some optical transmission systems are dominated by dispersion, which is not random noise, so we approach the problem differently.)

paolo.franzoi
50%
50%
paolo.franzoi,
User Rank: Light Sabre
12/5/2012 | 3:53:18 PM
re: The Ultimate Cable Modem


&nbsp;


Stevery,


Basically it is the SNR to be able to distinguish between coding points of the pattern, but you probably knew that.&nbsp; Only 3 ways to fix that:


1 - Transmit more power


2 - Transmit the same power over less spectrum


3 - Build a better receiver


Now, the reason I thought about noise is that having spent many years in DSL I know there are some non-periodioc noise environments (like Ring Trip in DSL) that are more fixable than others.&nbsp; The other thing that is assumed in Shannon (and NOT true in the real world) is that the channel is constant.&nbsp; In most cases, it can be considered constant over any given interval - except for say rain in wireless or water in copper plants or bending in fiber or etc.


Finally, all of these coding schemes have been around for a REALLY long time.&nbsp; Like since the 50s and originally came out of military comms.&nbsp; I know I worked on QAM and DPQSK leased line modems over 20 years ago (Go Go Omnimode!).&nbsp; So, the bit/hertz density is what it is.&nbsp; There have been some clever schemes at noise reduction (MIMO) and channel creation (OFDM) but there you go.&nbsp; All that has really happened is the cost of DSPs have fallen through the floor.&nbsp; Those leased line modem had 12V PMOS DSPs that we custom built as chips.


seven


&nbsp;

Stevery
50%
50%
Stevery,
User Rank: Light Beer
12/5/2012 | 3:53:16 PM
re: The Ultimate Cable Modem


Thanks for humoring me seven.


Let's fix the bandwidth for the rest of this discussion, eliminating case (2).&nbsp;


There is some point at which (3) won't help, because the receiver noise is sufficiently below the channel noise.&nbsp; (For example, the resistive losses in the cable give rise to thermal noise.&nbsp; More cable length = more resistive loss = more 4kTR = more noise.)


Also, practical issues of reflections, dispersion etc will limit how far you can get with (1).&nbsp; Although such items are not random noise, you can pretend it is from a system standpoint. (And yes, you can go compensate for it with a sufficiently complicated transmission system.&nbsp; Eliminate that possibility for the moment.)


So for the above restrictions, there is some implied SNR.&nbsp; In turn, this implies the maximum spectral efficiency (and maximum bitrate) for the channel.


I am wondering if a ballpark number for that maximum number is well known or not.


&nbsp;

rjmcmahon
50%
50%
rjmcmahon,
User Rank: Light Beer
12/5/2012 | 3:53:15 PM
re: The Ultimate Cable Modem


An analysis done 7+ years ago can be found here.

paolo.franzoi
50%
50%
paolo.franzoi,
User Rank: Light Sabre
12/5/2012 | 3:53:15 PM
re: The Ultimate Cable Modem


&nbsp;


So, let me continue your thought.&nbsp; There are 2 other practical issues:


1 - The channel noise itself - which is why fiber uber alles


2 - The cost and power of the DSP engines


So, I am thinking there is probably some reasonable bits per hertz efficiency that is possible if we use fiber (of some fixed length - not huge) as a channel to give us a low noise environment that we consider best in class.&nbsp; So, could you build a 10^10 QAM on a fiber environment?&nbsp; Good question - I have never done the math.&nbsp; The reason is that item 2 is generally the limit on fiber.&nbsp; The speeds obtainable are fundamentally limited by the semiconductor technology.


That is why I brought up the age of our coding schemes.&nbsp; I am not a math expert inventing these things, but there are people that are.&nbsp; I would have though that if there was the coding to beat all coding that these schemes would be implemented.&nbsp; Yet, we depend on coding schemes invented 50 years ago.&nbsp; I used to look for better coding schemes as a way to do longer reach DSL transmission with either better BER or higher bit rate.&nbsp; (I have seen the multipair MIMO stuff).&nbsp; I was looking around and the only one that made any sense to me at all was a Spread Spectrum Technique - but that never went anywhere.


So, QAM, QPSK, DPQSK, DMT etc all are varied on the costs to implement based on the transmission characteristics of the channel and the power/cost to implement the DSP.&nbsp; Is one scheme fundamentally more bits/hertz?&nbsp; I have not seen that and most schemes I have seen are really trying to either build better receivers or and more channel.


seven


&nbsp;

Stevery
50%
50%
Stevery,
User Rank: Light Beer
12/5/2012 | 3:53:14 PM
re: The Ultimate Cable Modem


The channel noise itself - which is why fiber uber alles


And so, we come to the conclusion that "The Ultimate Cable Modem" is to rip out your cable and install fiber! I couldn't agree more, but we're not going to pick up lots of hot marketing cable babes with that line.


I think the quest for the "The Ultimate Cable Modem" is much like the DSL evolution. No one believed that twisted-pair was the "best" means of transmission, but there was a huge installed base, so everyone was stuck with it for a while.&nbsp; Cable modems and FIOS continue to drive twisted pair into irrelevance, but still some DSLAMs get sold.


2 - The cost and power of the DSP engines


My take is different:&nbsp; DSP and silicon is cheap.&nbsp; Unfortunately, all of the mod schemes you list require phase, and coherent photonic receivers are not suitably priced for the consumer market.&nbsp; In fact, DPQSK (which is really 2xQAM4) is just beginning at the high end market (40+ Gbps) where the economics is plausible. Will it continue and improve to QAM8 etc?&nbsp; I think so.


@RJ:&nbsp; I'm afraid I don't see anything fundamental in that analysis.&nbsp; Maybe I'm not getting it.


The bigger issue:&nbsp; I assert that there is unused SNR in current cable systems.&nbsp; Please tell me why I should not go raise $10M to build The Ultimate Cable Modem, the TUCM model QAM65536.

rjmcmahon
50%
50%
rjmcmahon,
User Rank: Light Beer
12/5/2012 | 3:53:14 PM
re: The Ultimate Cable Modem


I should note that I don't totally buy into Frankston's rhetoric here but I do think that the technology case for a better cable MODEM may not align with a cable cos revenue incentives.&nbsp; VZ "copying" HFC with FiOS is a sign that the revenue and regulatory models are the challenge more than the technology.


&nbsp;


&nbsp;

rjmcmahon
50%
50%
rjmcmahon,
User Rank: Light Beer
12/5/2012 | 3:53:14 PM
re: The Ultimate Cable Modem


Stevery:


My interpretation of the analysis is that the cable MODEM is not the barrier to residential broadband from the cable companies.&nbsp; There would need to be more fiber for uplinks and CMTSs installed in the field.&nbsp; (Similar to what happened with remote DSLAMs and the PSTN)&nbsp; This is cost prohibitive on the operational side.&nbsp;&nbsp; And regardless, the primary revenue would get canabalized by 100Mbs+ links (video would go the route of mp3s.)


It kinda reminds me of Bob Frankston's one percent article written a few years back.&nbsp; An excerpt:


"They [VZ] haven't given up on ITV [interactive TV] &ndash; this time they are using fiber as the transport and calling it IPTV. The fiber they are installing for FIOS is really a cable TV plant disguised as a network. It is a Passive Optical Network (PON) designed as a distribution system from a head end to the terminals at each home though it does have capacity to send data back. A single fiber has the capacity for gigabits of traffic. There's so much capacity that they can simply allocate a portion of the capacity to emulating traditional Cable TV. The 15mbps they reserve for their Internet service is less than 1% of that capacity!


The big lesson of the Internet and personal computer is that it makes more sense to just deploy simple IP connectivity and then use standard digital technology to convert the IP video streams to analog video when necessary. While I might forgive the Telcos for neglecting the old-line telephony business, it's harder to understand why they are deploying technologies that are obsolete before they are deployed.


One reason may be that the tradition of &ldquo;CO (Central Office) Grade&rdquo; makes them very conservative and it seems a very safe choice when it is really a brittle choice. Even better for the carriers is that it maintains a distinction between video bits and Internet bits. The effect is to take the 99% of the bits &ldquo;off the table&rdquo; so they don't have to compete nor worry about efficiency.


The distinction between video bits and Internet bits maps nicely into the myths that define the FCC. The FCC treats the Internet as an information service rather than a fundamental technology.


This flies in the face of what we've learned since the regulations were put in place. The regulations date back to a time when the technology barely worked and every element of the system had to be precisely specified. It was very expensive and each signal had its own special characteristics &ndash; bandwidth, frequency, noise etc. IP technology is fundamentally different and allows us to have a single packet medium. All the packets are the same &ndash; video, audio, images, text &ndash; it doesn't matter!"


&nbsp;

paolo.franzoi
50%
50%
paolo.franzoi,
User Rank: Light Sabre
12/5/2012 | 3:53:11 PM
re: The Ultimate Cable Modem


&nbsp;


Stevery,


You and I probably have a different view of DSP cost reductions over time.&nbsp; As I said, I came originally out of the leased line and dial modem business - I worked for Racal-Milgo back in the day.&nbsp; We used to be able to charge (and the margins were very good) $1 per bit per second for Leased Line modems (lets say this is the early 80s).&nbsp; This is pretty much before off the shelf DSPs were available (TI 320 series being first iirc).&nbsp; These 2400 baud, 9600, 14400&nbsp;or 19200 bps modems were QAM and DPQSK based (they contained a very slow side channel FSK modem as well for management).&nbsp; So, I see DPQSK and QAM as having been implemented in DSPs for a LONG time.&nbsp; Telebit (iirc) built the Trailblazer 19.2 Kbps dial modem based on DMT in the mid-80s (maybe late 80s).&nbsp; The advent of Fax machines created a broad based modem market to get commericial silicon makers (back in the day Rockwell) to build devices.&nbsp; Derivatives of these eventually got built into dial modems with the advent of the web


The consumer devices (Dial Modems, Cable Modems, DSL Modems, ONTs) all have to trend to $100 per home or so - and they all have or become much cheaper (even indoor ONTs can be built for well under $100 for diplexer built system).


So, the ongoing drop in DSP prices has created the ability to run higher rates with more complicated DSP algorithms for about the same price.&nbsp; At the price of 40G fiber transport, the DSP costs are not the issue as you have pointed out.&nbsp; So, from my simple Access mentality I see that there is a progression based around the ability to build devices that sell for $100 (or thereabouts).


seven


&nbsp;

paolo.franzoi
50%
50%
paolo.franzoi,
User Rank: Light Sabre
12/5/2012 | 3:53:09 PM
re: The Ultimate Cable Modem


rj,


Sir, you have your history wrong on this.&nbsp; Prior to 1994 (WWW introduction), the vast bulk of all modems both dial and leased were bought by businesses.&nbsp; The bulk of dial modem "innovation" and churn occurred before this date - with only the introduction of 56K modems (both X2/K.56 followed by V.90/V.92) occurring later.&nbsp; But the progression of dial modems from 300 bps&nbsp;- 33.4 kpbs proceeded any significant consumer market.&nbsp; There was a significant Enterprise market with customers like Visa and Mastercard dominating the space.


seven


&nbsp;

<<   <   Page 2 / 3   >   >>
More Blogs from The Bauminator
Findings from university researchers and the WSJ could crash broadband's ARPU party as a growing number of subscribers upgrade to faster, more expensive high-speed Internet tiers.
The ad-driven video joint venture is attempting to carve out new deals around the world as impressions flatten out thanks to finite inventory and the rise of streaming competitors.
Combined company only will trail Disney, Comcast/NBCU and AT&T/WarnerMedia in that category as streaming battle intensifies.
Slated for launch in early September, the new offering will include a package of more than 50 US television channels, including 'faith-based,' premium and international channels.
Connections to AT&T and Dish harm Locast's position as a non-profit, do-gooder and add fuel to the arguments being set against Locast by the Big 4 US broadcasters.
Featured Video
Upcoming Live Events
September 17-19, 2019, Dallas, Texas
October 1-2, 2019, New Orleans, Louisiana
October 10, 2019, New York, New York
October 22, 2019, Los Angeles, CA
November 5, 2019, London, England
November 7, 2019, London, UK
November 14, 2019, Maritim Hotel, Berlin
December 3-5, 2019, Vienna, Austria
December 3, 2019, New York, New York
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events