[Discuss] data caps

[tmetro+blu at gmail.com (Tom Metro)]

Lets move on from DRM and GPL to another topic we all love - data caps!

The clips and commentary below became too long, so I'll provide a tl;dr
summary up top, and pose a question for discussion. Here's the premise:
data caps are not about solving network congestion, they're about
increasing revenues and staving off competition from other content
providers; data delivery has gotten increasingly more profitable for
ISPs as their delivery costs have dropped and their investment in
infrastructure has shrunk; the lack of competition permits this to happen.

Read further below if you want to see the articles that support the above.

Given this, if you were choosing a broadband provider, and you didn't
want to reward companies that follow these practices, who would you
pick? While you can currently avoid data caps by selecting a
business-class service, you're still rewarding the same companies with
your business, and what's to stop them from introducing caps later?

In the sub-$200/month price range, there doesn't seem to be an
alternative to cable and telco fiber, unless you are willing to slow
down to DSL speeds, or happen to be in one of the few areas where there
is a fixed wireless provider.

I recently ran across an old Linux Journal (May 2008) interview with Bob
Frankston, co-inventor of VisiCalc, where the topic of Internet
connectivity came up. The interviewer, Doc Searls, describes Frankston
as the Richard M. Stallman of connectivity. Here's a clip:

  BF: ...the network emerges out of our networking. Copper and radios
  are just a means we use. It's like the difference between driving and
  buying a ride from a railroad. We should have infrastructure rather
  than a choice of whose service we must purchase. DIY must be an
  option!

  DS: Why the railroad analogy?
  BF: Because we're still thinking in railroad terms. The FCC was partly
  an outgrowth of the ICC (Interstate Commerce Commission), which
  regulated railroads. Given the opportunity - which they were -
  railroad owners became infamous robber barons. How different is that
  from today when phone companies charge you for the contents of your
  freight cars, rather than just for using the track? Take SMS, for
  example. It's just data - a small number of bits using idle capacity.
  Yet an SMS bit costs millions of times more than a video bit.

  They can charge that because, like the railroad barons, they use their
  control of the infrastructure to force us to buy vintage services at
  arbitrary prices. These are phone and cable companies with railroad
  legacies. Not Internet companies.

  The importance of the Internet lies in the dynamic process by which a
  very simple design decision made in the 1970s has become the defining
  infrastructure for the world. It's what happens when you give billions
  of people the opportunity to create their own solutions and share
  them. The infrastructure of telecom is not the infrastructure of
  networking. We must not confuse the two. The infrastructure of telecom
  is about billing scarcity. The infrastructure of networking is DIY and
  connecting anything to anything.

Of course DIY connectivity is far easier said than done. For years there
have been various projects to create rooftop mesh networks and the like,
mostly in urban areas. Coverage is still small. A few people on this
list work on such networks, but how many subscribers here are in range
of them? Even if you are, do they offer the speed and reliability to
support a small business? If not, then they aren't a practical
alternative yet.


The remainder below are the articles discussing data caps.

The Open Technology Institute recently released a report on data caps
titled "Capping the Nation's Broadband Future?" Here are some highlights:

http://www.newamerica.net/publications/policy/capping_the_nation_s_broadband_future

  ...while new services and applications require more data, most major
  Internet service and mobile providers in the United States are moving
  in the opposite direction: discouraging Internet usage by implementing
  more restrictive and costly limitations or data caps.
  ...
  This combination of caps and higher prices also applies to services
  like text messaging. Even though "the cost a carrier incurs by
  transmitting an SMS message has not increased in recent
  years, "carriers have continued raising prices and imposing
  limits. As with data caps, the prices charged to consumers do not
  correspond with the costs for carriers.

  ISPs often claim that caps are necessary to curb "excessive use" and
  only affect a small fraction of users. Although some providers are
  reexamining their data caps policies, many of the limits imposed
  several years ago have largely remained static, even as typical
  household bandwidth consumption has substantially increased.
  ...
  Data caps encourage a climate of scarcity in an increasingly
  data-driven world. Broadband appears to be one of few industries that
  seek to discourage their customers from consuming more of their
  product. Thus, even as the economic and engineering rationale for data
  caps on wireline broadband does not hold up given the declining costs
  of providing service and rapid technological advancement, the
  proliferation of data caps is increasing.

  The trend is driven in large part by a woefully uncompetitive market
  that allows the nation's largest providers to generate enormous
  profits as well as protect legacy business models from new services
  and innovators.
  ...
  Costs to provide broadband service for the biggest ISPs are largely
  declining even as they add new subscribers to their networks and
  overall traffic is increasing. Moreover, the use of monthly caps are
  largely ineffective at limiting network congestion that is often
  limited to specific times and places.

  Lastly, a number of ISPs' capital expenditures have declined over the
  past several years, even as their profits on broadband service have
  substantially increased.

  The evidence suggests that the imposition of costly or punitive data
  caps are largely a business decision by ISPs to boost revenues per
  subscriber or protect legacy services in a market where consumer have
  few choices and where consumers only switch services on a very limited
  basis.


The paper goes on to examine how the costs for Internet backbone peering
have been steadily decreasing, and how overall service delivery costs
have declined, as shown by ISP's own public disclosures to their
investors. (In the case of Comcast, it cost them $11/user/quarter in
2007 to provide high-speed Internet, while in 2010 that had dropped to
$7/user/quarter.) The paper also cites Google's statements about their
Kansas City fiber project, "There's no need to be limited...There's no
need for caps."

What irks most technical people is the lie that data caps are there to
address network congestion:

  Many providers claim that data caps are necessary as their networks
  are becoming increasingly congested and unable to handle heavy usage.
  Yet monthly data limits are not the most efficient tool to address
  issues surrounding network capacity.

  The main challenge for network engineers is how to deal with demand
  during peak hours. Prime time evening hours see the most demand for
  bandwidth on the network, and managing this traffic can be a
  challenge. But an individual's total aggregate data consumption over
  the course of a month is not a concern from an engineering point of
  view. When an individual consumes data matters more than how much data
  he or she uses.
  ...
  A flat cap unnecessarily limits use even when the network is not
  congested. Indeed, even Comcast acknowledged this in a letter to the
  FCC. ...this cap "does not address the issue of network congestion,
  which results from traffic levels that vary from minute to minute."

Further, the report says that a logical way to address congestion in the
long term would be to invest further in infrastructure, but public
disclosures from ISPs indicate the opposite.

  ...a review of the publicly available financial document for some of
  the largest ISPs in the country shows a decline in capital
  expenditures--the costs associated with building, upgrading and
  maintaining a network, such as construction, repairs, and equipment
  purchases--for their wireline networks.Many ISPs are spending less
  money on capital expenditures now, both as a ratio to revenue but also
  even in raw dollars,than they have in years past.

  While some cost decreases can be explained by declines in hardware and
  equipment costs, these trends suggest that broadband providers are
  content to maintain the status quo and reap these efficiencies as a
  bonus rather than an opportunity to increase investment.
  ...
  Some estimate that cable broadband providers enjoy gross margins as
  high as 95 percent, an exceptionally high rate of revenue relative to
  the supposed costs associated with offering the service.[22]For these
  companies, selling broadband packages even to the heaviest users is
  still quite profitable.

The above understates the situation for telcos, as they've admitted to
boosting the backbone infrastructure for their wireless networks, while
recording it on the books as improvements to their wired operations.
"...Verizon's spending on improving its DSL products or expanding FiOS
are even less than they appear on financial disclosures..." (Even
Comcast has a department dedicated to selling wired backbone services
("backhaul") to wireless carrier, and that infrastructure probably goes
onto the same line item as their retail networks.)

So are caps just a way to keep users subscribing to cable TV?

  A number of industry observers speculate that data caps are less about
  network management and more about discouraging consumers from
  accessing content online which they traditionally consumed
  offline. As suggested by the Department of Justice probe of data cap
  policies, caps are a useful tool for cable companies to protect their
  legacy video services.

Clearly these companies have a conflict of interest.

The report goes on to discuss the situation with data caps with cell
networks, which I won't summarize here.

The report concludes that data caps are not technically justified, and
exist to increase profits and act as barriers to content competitors.

  Yet as average consumer usage approaches the caps, the carriers have
  offered little if any data that supports their claims that the caps
  they have implemented are necessary. At the same time, broadband
  service has become an increasingly profitable business. In May 2012,
  Comcast exceeded profit and sales estimates and saw its net income
  rise 30 percent and average revenue per user increase by 7.8 percent.
  ...
  At the same time, overall growth in the number of broadband
  subscribers is leveling off, providing a significant incentive for
  both wireline and mobile providers to create scarcity and find
  additional revenue streams from their existing customer base. Charging
  for data usage through caps offers a lucrative means to do that, and
  providers are largely able to do so because of the lack of competition
  in the U.S.

And as many other reports have noted, the US is in worse shape than may
other countries:

  ...in addition to fewer choices, higher prices and slower speeds, ISPs
  in the U.S. trended toward tacking on additional fees and data caps
  far more frequently than their international peers. In competitive
  markets like Paris and Hong Kong, for example, few ISPs impose data
  caps on their wireline networks.

The solution prescribed by the report simply is the usual of creating
policies that "encourage competition."

That's really insufficient, if most conditions remain as-is. We'll just
get more of this:

http://www.bloomberg.com/news/2012-12-26/how-at-t-and-verizon-manipulate-your-smartphone.html

  The two kinds of Internet-access carriers, wired and wireless, have
  found they can operate without competing with each other. The cable
  industry and AT&T- Verizon have divided up the world much as Comcast
  and Time Warner did; only instead of, "You take Philadelphia, I'll
  take Minneapolis," it's, "You take wired, I'll take wireless."

  At the end of 2011, the two industries even agreed to market each
  other's services. Comcast and Time Warner Cable will bundle Verizon
  Wireless services with their own, and by 2015 the cable companies will
  have the option of selling mobile services under their own brands.
  ...
  Comcast and FiOS overlap in only 15 percent of Comcast's physical
  market; Time Warner and FiOS overlap in just 11 percent of Time
  Warner's. By cooperating, Verizon Wireless implicitly promises not to
  spread FiOS service any farther, and Comcast and Time Warner promise
  to stay out of the wireless business.
  ...
  Now, the communications industry is at a point of equipoise. Each of
  the major actors is too big for any of the others to swallow or crush.
  Profits are climbing, allowing the companies to pay ever-higher
  dividends. Cash is piling up; investment in infrastructure is down,
  because there is no competitive pressure to increase it.

And this suggests changes are even less likely to be proposed:

http://www.dslreports.com/shownews/FCC-Hires-New-Cap-Loving-Chief-Economist-122566

  The FCC has hired a new chief economist with a history of cheerleading
  broadband usage caps for the cable industry. According to the FCC,
  they've hired Steven Wildman, an economist and professor at Michigan
  State University...

  [He] just got done penning a National Cable & Telecommunications
  Association paper supporting the industry's use of punitive caps and
  costly per-byte overages. "...The effects of well-designed
  [usage-based pricing] plans on consumers are likely to be beneficial,
  as are the effects of UBP on investments in the broadband
  infrastructure," insisted Wildman in the cable industry sponsored
  paper (pdf).
  ...
  Like with most of these industry-funded studies, Wildman ignores the
  fact that the lack of competition drives the creation of punitive caps
  and overages in the first place, and that few if any of these plans
  are "well designed." Most of what we've seen implemented by carriers
  so far isn't really usage-based pricing, it's flat rate pricing with
  caps and per-byte penalties layered on top, resulting in higher prices
  for everybody.
  ...
  Wildman's hiring is troubling for an agency that has already utterly
  refused to seriously address the anti-competitive impact of usage caps
  and high overage pricing. Genachowski's positions on caps have waffled
  depending on what lobbyists or industry leaders he's talking to, and
  that's just part of a larger FCC problem with failing to address
  competition.


Public Knowledge, a site that frequently reports on these matters, also
weighs in on data caps:

http://www.publicknowledge.org/blog/question-core-data-caps-debate

  ...as we have argued, data caps do not resolve congestion, are
  confusing to consumers, and lend themselves to unfair and
  anticompetitive behavior.

They take a closer look at Steven Wildman's report mentioned above.

  ...it is a promising sign that a recent study published by the
  National Cable & Telecommunications Association (NCTA) and co-authored
  by Steven S. Wildman, the new Chief Economist of the FCC, moves
  beyond some of the previous rhetoric and takes a significant step
  towards focusing the debate on real areas of conflict.

  Unfortunately, it stops short of recognizing a critical distinction in
  understanding the heart of the disagreement.

Public Knowledge sees it as a good sign that the report doesn't say that
data caps or overage charges addressed congestion. Instead the report
says that usage-based pricing lets "people who value something more pay
more for it and allowing people who value something less pay less for
it." PK points out that having different price tiers is nothing new -
most ISPs will sell you faster speeds for more money - the problem is
tying costs to bits, and they say the report doesn't address that.

  Public Knowledge's position is that data caps and usage-based pricing
  is a type of price discrimination that is especially susceptible to
  anti-consumer manipulation by ISPs and can suppress activities that we
  generally encourage.  That is why they warrant attention.

(That's discrimination in the differentiation sense, not the repression
sense, so not necessarily a bad thing.)

PK argues that the feedback loop to end-users is much clearer for speed
than it is with data caps. With speed, you get immediate feedback when
things are slow. With caps, you find out at the end of the month that
you have a pile of overage charges.

I tend to think this is a minor side point. The far more import problem
with charging for a fixed quantity of bits is that it has no actual
connection with operating costs. It's a purely artificial constraint.


Another article comparing the US broadband to other countries starts off
with a familiar story of a community fiber project and how the incumbent
providers tried to legislate it out of existence:

U.S. Internet Users Pay More for Slower Service
http://www.bloomberg.com/news/2012-12-27/u-s-internet-users-pay-more-for-slower-service.html

  In 2004, the Lafayette [Louisiana] utilities system decided to provide
  a fiber-to-the-home service. The new network, called LUS Fiber, would
  give everyone in Lafayette a very fast Internet connection...
  ...
  Push-back from the local telephone company, BellSouth Corp., and the
  local cable company, Cox Communications Inc., was immediate. They
  tried to get laws passed to stop the network, sued the city, even
  forced the town to hold a referendum on the project -- in which the
  people voted 62 percent in favor. Finally, in February 2007, after
  five civil lawsuits, the Louisiana Supreme Court voted, 7-0, to allow
  the network.

  From 2007 to mid-2011, people living in Lafayette saved $5.7 million
  on telecommunications services.

Other communities weren't so lucky...

  ...in 2011, six Time Warner lobbyists persuaded the North Carolina
  legislature to pass a "level playing field" bill making it impossible
  for cities in that state to create their own high-speed Internet
  access networks. ... Eighteen other states have laws that make it
  extremely difficult or impossible for cities to provide this service
  to their residents.

The article then mentions the Google fiber project, almost as if it is
in the same realm as a community fiber initiative. What does that say
when an industry is so bad that competition from the 7th largest (2012
Forbes 500 by market cap) company is viewed as a disruptive underdog.

The article makes the case for the community approach by likening it to
electricity:

  Internet access, like electricity, is crucial to the economic and
  social health of the country. Electricity, however, is provided by
  largely reliable, taxpayer-supported entities, and no one seems to
  think the country would be better off if a purely private, wholly
  deregulated operator were in charge.
  ...
  By the mid-1920s, 15 holding companies controlled 85 percent of the
  nation's electricity distribution, and the Federal Trade Commission
  found that the power trusts routinely gouged consumers.

  In response, thousands of communities formed their own electrical
  utilities. Now more than 2,000 U.S. communities, including Los
  Angeles, San Antonio and Seattle provide their own power. And
  electricity is a regulated public utility.

  Why don't Americans apply this same thinking to communications?

Seems like sound reasoning, except *only* 2000 communities? The idea
doesn't really scale. The majority of US communities would be unable,
and likely woefully inefficient at implementing a fiber infrastructure,
redundantly re-inventing the management overhead in each and every town.

The article goes on to talk abut how the FCC "National Broadband Plan"
is flawed in that it adopts the cable companies' asymmetric broadband
model, where broadband is primarily used to receive entertainment, not
to run a business, and how the speed targets are set too low (4 Mbps
down). In contrast, the "South Korean government announced a plan to
install 1 gigabit per second of symmetric fiber data access in every
home by 2012."

Other countries seem to consider broadband connectivity to be vital to
their global competitiveness. The US tends to get bogged down with its
internal political fights.

The article has some interesting stats:

  ...the U.S. is rapidly losing the global race for high-speed
  connectivity, as fewer than 8 percent of households have fiber
  service. And almost 30 percent of the country still isn't connected to
  the Internet at all.
  ...
  How much would it cost to bring fiber to the homes of all Americans?
  Corning Inc. (GLW), the American glass manufacturer, and others have
  estimated that it would take between $50 billion and $90 billion.

That price seems cheap given the sort of return you'd get on it the
investment.

So what did we pay for the big dig? One site says $22B once you add the
interest. (Doesn't mention inflation.) So for about 1/3 the cost of
national broadband we made traffic run a bit smoother in one tiny spot.

Or another point if comparison: supposedly the "fiscal cliff" deal
included $60B in corporate tax breaks.


  To make this happen, though, the U.S. needs to move to a utility
  model, based on the assumption that all Americans require fiber-optic
  Internet access at reasonable prices.

Utility model - yes, but not necessarily a government owned and ran
utility. What we seem to be missing is a government defined framework
for a "last mile" company, which would be carefully designed to align
the incentives with serving the consumers of a last mile service.

If you really expect to reach near 100% coverage of households,
inevitably you would still need government funding for areas where the
population is too sparse. That still doesn't necessarily mean government
operated fiber. It could be in the form of a zero interest loan to a
provider for the up-front installation costs, or the government could
pay for the installation, own the last mile, and lease it to a provider
to operate and maintain it.

There are still some glaring holes in this approach. It doesn't achieve,
or even encourage, competition in the last mile itself, which could
stagnate, if it is too tightly regulated.

Ultimately, any such scheme would require displacing the telco and cable
companies in their exclusive grip over the last mile, and lower the
barriers for content competitors, which means they'd funnel all their
profits into lobbying against any laws that set up new last mile rules.

 -Tom

-- 
Tom Metro
Venture Logic, Newton, MA, USA
"Enterprise solutions through open source."
Professional Profile: http://tmetro.venturelogic.com/