Fast Net News, DSL Prime, DOCSIS Report

RAND and other cold warriors developed the Delphi method of asking experts for opinions, then recirculating the comments and asking the original panel to rethink. The goal is for the discussion to lead to a "carefully considered expert consensus." The result is what Herman Kahn called a "surprise-free" projection. No prediction is certain, of course; Herman's own work Towards The Year 2000 had some right and many wrong. With time and funding, I'd add a slew of scenarios. While a Delphi survey has the trappings of science, it's just an organized collection of opinion. The same is true of almost any social or policy prediction, of course.

I will program this to a web form when I can. For now, simply cut and paste the below into an email to daveb **at** dslprime.com either directly or via the gmail account broadbanddelphi@gmail.com password kahnherman. Put answers under "Your thoughts." Questions 7,8,10, and 11 aren't easy to answer with a number, so please place an "x" according to the form and add narrative if you like.

Here's the questions and comments:

1) Availability, modest speed: more

Total DSL + Cable + Fiber + Fixed wireless, modest speed generally 3 meg or above.
Current FCC figures  3-6M are "unserved," or about 5% of homes. (Broadband plan slides. September 2009)

2012-2013

First thoughts: The "unserved" will be reduced to 2-3%, unless policy changes, with a few at the edge perhaps at speeds slower than 3 megabits.

Other comments:

The rate of "unserved" will be more rapidly reduced because advances in DSL technology (DSM, bonding) will allow reaching many of them.
1%-2% of households
1% I think broadband will be as available as electricity by then.
5% because satellite will serve most of the unreachable so no justification for trying to get the last 5% on wireline

Your Thoughts:

2014-2016

First thoughts: The unserved will be reduced to 1-2%%.

Other comments:

less than 1%
4%.  see above about satellite. the wireline edge will be pushed out very slowly

Your Thoughts:

2020

First thoughts: 0.5% - 1.5%

Other comments:

0.5%
A certain number of homes are brutally expensive to reach and best left to improved satellite.
3%  just like telephone penetration.  But satellite will be okay.

Your Thoughts:

2) Higher speed availability. Generally 50 meg down or above via DOCSIS 3.0 or Fiber

Current figure, my estimate Fiber 15M, DOCSIS 3.0 35-40M. Because there is some overlap, I'd estimate the total at 45-50M, just under half the U.S. homes.
2012-2013

First thoughts: 88-92% will be able to get 50 meg 95+% of the time

Other comments:

Network decisions behind the cable head-end or DSLAM will be the limiting factor.   Most of the 92% that can get cable modems will be upgraded to DOCSIS 3.0, and a handful of other places will be reached by rural fiber, etc.
Advances in DSL (bonding, DSM) will enable as many as 30% of the DSL lines to achieve downstream faster than 50 megabits.
90%-95%
90%
90% available, not taken
[C] guesses all major cablecos will fully upgrade to DOCSIS 3.0 within two years, maybe faster.  They see a major technological advantage over telcos here and are going to exploit it.  I don't know about the smaller companies.  Some are in very good shape, but there are still some small systems out there and I don't know where they will get the extra channels for bonding.

Your Thoughts:

2014-2016

First thought 90-95%

Other comments:

Fiber to nodes/cabinets will increase, but fiber to the home will wane as people understand it is not necessary to more than accommodate the bandwidth need.
Some of the 4% of homes that can get cable TV but not data will be upgraded
95%
still 90% because demand will be too thin in the difficult-to-serve areas to justify the expense.  However, it moves to 95% if a broadband Universal Service Fund subsidizes the upgrades

Your Thoughts:

2020

First thoughts: Few

Other comments:

97%-98%
98% with satellite and 4G
91% un-subsidized, 95% subsidized

Your Thoughts:

3) Advanced wireless LTE or WiMAX at 4-10 meg down

2012-2013

First thoughts: 90-95%

Other comments:

Verizon's 92% sets the stage.
[C] would expect 90 - 95% coverage, as well.  While Verizon has said it plans to have nationwide coverage by 2014, it has not clarified what that means.  Perhaps in terms of POPs it will meet its goal, but I don't believe in terms of capacity it will be there.  By 2014, it will have rolled out its initial nationwide footprint.
Severe congestion will arise as various apps (think iPhone for how many apps can eat up bandwidth), forcing smaller-cell solution.  The leaders in recognizing this and how they do it will eventually dominate the telecom world.
95%
90%  of population, not square miles since  we have a lot of uninhabited areas in this country.  This is where the broadband stimulus will show results if enough mid-mile projects are supported by BIP/BTOP so that WISPs can deploy WiMax in the unserved areas.
[C] would expect 90 - 95% coverage, as well.  While Verizon has said it plans to have nationwide coverage by 2014, it has not clarified what that means.  Perhaps in terms of POPs it will meet its goal, but I don't believe in terms of capacity it will be there.  By 2014, it will have rolled out its initial nationwide footprint.

Your Thoughts:

2014-2016

First thoughts: 95-99%

Other comments:

Carrier plans are aggressive.
[C] would say 95%.  There are regions of the country that I don't believe operators will get an ROI on their investment and will not go afte
98%-99%
90%  satellite makes the business case for extraordinarily expensive deployment of terrestrial wireless uneconomic and unnecessary

Your Thoughts:

2020

First thoughts 95-99%

Other comments:

99%
[C] will stick with 95%.  Between 2015 and 2020, LTE investment will be targeted towards matching capacity with the demand, not increasing the footprint.
100%
91%  satellite is always more economic than wireless and is adequate operationally

Your Thoughts:

4) Take rate, including high speed mobile

Current figure, about 65% of U.S. homes

General comments:

[C] wanted to react to the question about uptake. I worry that the FCC will continue to look at this as a technology driven "problem"--add some new technology and shift the answer. According to Pew, 22% of Americans say that they do not use the Internet. The contributing factors seem to be age (older folks are much less likely to be users), education and income levels. To put it plainly, as the current population of elderly exit the population, there will be some natural increase in overall usage. But aside from attacking the issues of cost through technical innovation, most of the issues here are social, and if we (as a country) think we should have a policy of increasing this number, it has to require non-technical intervention.
[C] one way to explain the higher percent of Asian households that have broadband is to note that in Asia many elderly live with their children, so even if they do not use the Internet, they do not show up as a non-using household. The many single-person elderly dwelling units we have in the US are not as common in the Asian housing stock.
[C] think for all these "lower-uptake" populations, there is a trend over time that they find value and become users. But I think the last 15% will be persistent and hard to move. Today, even among high-income people, the uptake is 85%, and about the same for those with college education.  Since the current population of non-users is about 22%, that would suggest that after another 7 percentage points come online, we will find the uptake appears to saturate, and moves upward only very slowly.
Korea has been stuck at about 80% household penetration for years, and lots of things about Korean Internet adoption looks surprisingly like ours (I'll forward you a new report by Korean ministry that's very interesting).  We never got above 96% household penetration for telephones, despite extensive subsidy programs.
From current trends and info like that in the above para, my guess is that we will get to 80% household penetration within a few years and then start on a very slow growth path.  The only way we can do better is through lifeline/linkup programs for low-income.  80% hhld penetration will mean nearly 100% for everyone making at least $75k a year, and a steady falloff below that.  As you said, the only thing that will help is lower prices (or, equivalently, some kind of subsidy that makes the effective price lower).  "Digital literacy" is unlikely to do very much in the grand scheme of things.

2012-2013

First thoughts: Originally 75-85%, including high speed mobile. Several early responses were lower at 70%

Other comments:

[C] would guess 70%.  I believe penetration of mobile phones will go up, but the majority of new phones sold are not to new wireless customers, but to current wireless customers upgrading phones.
70%
80%
80% with substantial increase in wireless broadband and wireline broadband penetration leveling off

Your Thoughts:

2014-2016

First thoughts 80-90%,

Other comments:

Lower, perhaps 75-80%.  Many people are dropping landlines for mobile only. However, I would guess people going mobile only were not customers with only a landline that they churned for their mobile phone.
All the growth will come from mobiles and 5-25% of homes will drop landline broadband
80%
90%
85% I agree that wireless "cutting the cord" will apply to lower speed broadband so wireline penetration will decline, probably substantially , as it has  for basic wireline telephone lines.

Your Thoughts:

2020

First thoughts: 90%, with high speed data included in most mobile phones

Other comments

95%+
90%
90% will use broadband of some sort. "Take rate" implies an affirmative consumer choice so high speed it may be included in most mobile phones, but not everyone will purchase "high speed".  I presume that tiered services will be offered with 25% selecting a "lower speed" and a few will still be "voice only" even if the voice is VoIP over a data channel.

Your Thoughts:

5) Traffic growth, per broadband subscriber per month (?Exaflood vs. Moore's Law)

Current: Per Professor Odlyzko and Cisco, we have pretty reliable data the rate has been 25-45% traffic growth each year. That seems to have slightly dropped in 2008.

General comments:

First thoughts: Continuing between 25-45%, with uncertainty about the growth from web video.
So far, neither YouTube nor Hulu has resulted in an increase in the rate of growth.
[C] think data would only increase over time, particularly on the wireless side, as 4G handsets become prevalent over the next 5 years or so.  I am also a believer in video disintermediation (Hulu, etc.) and I think that probably reaches a tipping point in the next 4-6 years.
[C] guesses there is probably going to be a fairly rapid increase above previous trend as more video goes online.  But after some period of time that rate of increase will fall back to the previous level.  I think exaflood is baloney cooked up by ATT and adopted by Cisco.
The other thing to think about here is whether we move away from all-you-can-eat pricing to metered pricing.  Set aside the question of whether such pricing would harm consumers.  If adopted widely it wold change the patterns of traffic growth.

2012-2013

Other comments:

35%
35% HD and 3D
40% per annum as video ramps up before volume caps and various pricing plans start to throttle video

Your Thoughts:

2014-2016

Other comments:

30%
It's always hard to predict creativity in bandwidth use, but said use never seems to go down.  Somewhere in here I’d expect some significant applications in cloud services to many customers  
The internet is maturing: so 25% per annum as network management practices become commonplace, reducing demand, the laws of very large numbers take over, and human beings reach their limit of absorption of voice, data and video...no more hours in the day.

Your Thoughts:

2020

Other comments:

20%
Look back 10 years from 2009 – what things did not exist back then that do now that surprised us all?  (iPhones and blackberries did not exist that I know of, DSL was stuck at “G.lite” attempts at a few hundred kilobits in Korea and we all were still on dial-up, maybe 1.5Mbps dream in JPC in USA, and Ipod did not exist nor did anyone anticipate flash challenging disks, unbundled competition was a rage and everyone thought telecom stocks were a good investment and the polar ice cap still existed to Canada and Siberia.  VCR’s gave way to DVD players and TiVo, while the CD after freshly replacing the record is now largely being replaced by flash memories.)  The equivalent would probably be that we’re all using some kind of very power personal communicator as if our lives depended on it, communicating in some form easily with anyone anywhere in the world instantaneously, services that border on us all having a personal assistant to lead us through our day, seamless purchasing, tracking.  Orwell was not that far off, just how long it would take.  Kind of scary actually.

Your thoughts:

6) Cost per bit of carrying traffic

Current: Traffic costs go down as Moore's Law improves the performance of routers and other equipment. Since about 2002, the traffic carrying cost has gone down at a rate of 25-45%, about the same as the amount of traffic has gone up. The happy result has been a roughly fixed cost per month per customer of carrying traffic.

First thoughts: Continuing between 25-45%

2012-2013

Other comments: To come

Your Thoughts:

2014-2016

Other comments: To come

Your Thoughts:

2020

Other comments: To come

7) Wireline competition

To simply things, I'm asking you to chose one of three options, writing in something different if appropriate.
Duopoly dominates, with limited additional choices most places
Stronger competition with more competitors, with a strong 3rd or 4th carrier emerging
Weaker competition, as cable pulls far ahead in some areas or telcos reassert their position. (Plausible, if line loss continues at >7%/year, but a minority position.)

First thoughts: Status quo continues, with differing thoughts on 5-10 years out

Other comments:

None of the above.  I see strong, perhaps more malevolent, Bell-entity emerging that comes from non-Telco space and buys all the com assets into a huge service- and connection- providing entity.  Kind of the “Gates” of personal computing instead of personal computers. I don’t know who.
National boundaries will be come less and less important.
Unless there is benign (non-political) and capable (remember Katrina and
post-war Iraq) government regulation, (highly unlikely on both poits) the
duopoly will continue. The two major telcos have the money and the political
clout to remain a player.  
Telco/cableco competition will be vigorous, though cable has an enormous advantage where it competes with qwest, embarq, and, to a lesser extent, att.
[C} is bullish on wireless competition, though.  I do not believe that even LTE will be a full competitor to wireline.  However, for the foreseeable future the majority of things people do can be easily handled by even 3G (a robust 3G, which maybe VZ has but T doesn't seem to yet).  That will be enough to impose both some price pressure and incentive for wireline to invest in QOS (maybe that means speed, but it might mean something else) as a means of differentiating themselves.  I wish I knew the degree of substitution between wireline & wireless, but I don't.

2012-2013

Other comments:

Still enough growth from new subscribers, new applications and refined pricing plans (aka, tiering and capping) to keep the duopolists reasonably happy
Satellite will be a stronger competitor within two years.  According to Wildblue, their new satellite will be capable of delivering 10-15 mbps to a user.  I heard some discussion of technology which would reduce latency by half.

Your thoughts:
___ Duoply, similar to today
___ More competitors, or substantially stronger competition among duopoly
___ Somewhat weaker competition
Explanation ____________________

2014-2016

Other comments:

Cable begins to slip as AT&T begins to focus on FTTH
As the internet matures, price competition breaks out as the only path to "growth" is to displace the other duopolist.

Your thoughts:
___ Duoply, similar to today
___ More competitors, or substantially stronger competition among duopoly
___ Somewhat weaker competition
Explanation ____________________

2020

Other comments:

Cable MSOs are largely content providers and Bells provide dominant wireline pipe into the home in 70% of the country
Wireline infrastructure becomes a regulated rate-of-return monopoly in the many markets that can't economically sustain multiple duplicative networks.  Cable is the survivor in some markets, telco in others and the winner is required to wholesale capacity to the loser and all other retail service providers.  Remember, by 2020 wireless has supplanted a lot of low cost, high margin slower speed wireline broadband so cablecos and telcos suffer from fixed costs trying to support reduced numbers of lines and lower prices.  The fate of a commodity.

Your thoughts:
___ Duoply, similar to today
___ More competitors, or substantially stronger competition among duopoly
___ Somewhat weaker competition
Explanation ____________________

8) Mobile broadband conversion

Mobile, at 4-10 megabits, is a partial substitute for wireline. There will be some conversion, with some people dropping their wired connections. The question is how many will go mobile only. Again, three options: Modest > 10%; moderate 10%-20%; large > 20%.

First thoughts: To come

2012-2013

Comment tally (x indicates choice):
modest > 10% xxx
moderate > 10-20% x
large > 20% x

Your thoughts:
modest > 10%
moderate > 10-20%.
large > 20%

2014-2016

Comment tally (x indicates choice):
modest > 10%
moderate > 10-20% xxxx
large > 20% x

Your thoughts:
modest > 10%
moderate > 10-20%.
large > 20%

2020

Comment tally (x indicates choice):
modest > 10%
moderate > 10-20%.
large > 20% xxxx
Surely wireless broadband and mobile will be universally available at speeds that will do to wireline broadband what wireless phone service has done to wireline phone service.
Virtually everyone goes mobile as part of the huge service-providing entity (that will use the fixed-line to get to the cell sites).

Your thoughts:
modest > 10%
moderate > 10-20%.
large > 20%

9) Improved upstream, DOCSIS 3.0 or wireless technology over 20 megabits upstream.

I'm leaving out upstream for fiber because that can be derived from the total number of fiber customers. Nearly all can receive over twenty megabits or be inexpensively upgraded. However, the DOCSIS 3.0 upstream is behind the downstream, with the first commercial deployments in 2010. Only Comcast and Wide Open West has announced plans for upstream, and some cablecos may delay the upgrade. I do not know of any commercial wireless technology that will bring over 20 megabits upstream to homes, but some may be available over the decade.

2012-2013

First thoughts: 60%

Other comments:

Comcast and some other cablecos will upgrade quickly (2010-2102) while others will delay. Until they see the results of early deployments, I don't believe most companies will set their plans. On wireless, I would be surprised if there will be many with fast upstreams.

Your Thoughts:

2014-2016

First thoughts: 75%

Other comments:

There are many proposals that are not close to commercialization and might be important later in the decade.

Your Thoughts:

2020

First thoughts: 90%

Other comments: To come

Your Thoughts:

10) Price, Average revenue per user, for most popular service

Current: Depending on the way this is measured. the current figure is between $32 & $50. Cable is generally higher than telco. There is no standard single measure, whether sold standalone or part of a bundle.

Please answer as: about the same; slightly lower, slightly higher, significantly lower, significantly higher.

2012-2013

Comment tally (marked by x as comments come in)

about the same
slightly lower
slightly higher
significantly lower
significantly higher.

Other comments: To come

Your thoughts:
about the same
slightly lower
slightly higher
significantly lower
significantly higher.


2014-2016
Comment tally (marked by x as comments come in)
about the same
slightly lower
slightly higher
significantly lower
significantly higher.

Other comments: To come

Your thoughts:
about the same
slightly lower
slightly higher
significantly lower
significantly higher.

2020

Comment tally (marked by x as comments come in)

about the same
slightly lower
slightly higher
significantly lower
significantly higher.

Other comments: To come

Your thoughts:
about the same
slightly lower
slightly higher
significantly lower
significantly higher.

11) Broadband capital spending

Current: Most companies only report total capital spending, of which broadband is typically a small fraction. FiOS, for example, is less than 20% of Verizon capex. AT&T in 2009 is cutting overall capex by about 10%, but is cutting U-Verse deployment by one-third.
My expectation is the trend will be down significantly between 2010 and 2012. Verizon, Comcast, and Time Warner have all announced they intend to reduce capex in 2010-2011. I believe the deployment of 4G wireless will slow in 2014-2015, but that's harder to predict.
Please answer as: about the same; slightly lower, slightly higher, significantly lower, significantly higher.

2012-2013

Comment tally (marked by x as comments come in)

about the same
slightly lower
slightly higher
significantly lower
significantly higher.
Other comments:
Verizon Comcast, and Time Warner have indicated they intend to reduce spending in 2010-2011

Your thoughts:
about the same
slightly lower
slightly higher
significantly lower
significantly higher.

2014-2016

Comment tally (marked by x as comments come in)

about the same
slightly lower
slightly higher
significantly lower
significantly higher.

Other comments: To come

Your thoughts:
about the same
slightly lower
slightly higher
significantly lower
significantly higher.

2020

Comment tally (marked by x as comments come in)

about the same
slightly lower
slightly higher
significantly lower
significantly higher.

Other comments: To come

Your thoughts:
about the same
slightly lower
slightly higher
significantly lower
significantly higher.

I'll follow up if this proves interesting. I'm only doing U.S. for now, although I'm happy to work with people elsewhere. These were the questions the FCC people thought most important.

If you want to be anonymous, log on to the Gmail account broadbanddelphi@gmail.com password kahnherman Cut and paste the questions please to help me stay organized. As far as I know, no one short of NSA with a subpoena will be able to trace you, and they aren't likely to be interested. If you trust I will anonymize you, just reply to the email.

Improvements welcome. Ask me any questions that might help.