Professional Documents
Culture Documents
Mr. S. M. files this Complaint to the Competition Bureau under sections 1.1, 76, 78 and
90.1 of the Competition Act. This complaint refers specifically to the recent
near-simultaneous admission by Shaw and Telus executives that they plan to
introduce usage based billing on their customers in the coming months. This
complaint also refers generally to anti-competitive application of usage based
billing by all other incumbent ISPs in the Canadian internet service market. Mr.
S. M. is a private citizen who has observed a number of CRTC proceedings
involving internet throttling and usage based billing (UBB) by incumbent
internet service providers in Canada since March 2008, including those
proceedings leading to leading to Telecom Decision 2010-802. He has observed
these because coordinated restriction of supply, price fixing, other anti-
competitive, collusive, and cartel-like application of billing practices among
dominant firms in the internet service provision sector in Canada negatively
affect S. M.'s ability to choose an ISP that can provide globally competitive
access to the internet. Furthermore the collusive behaviour of the incumbent
ISPs have the effect of degrading the ability of competing firms in the video-on-
demand business to provide him with the full quality of their services, by
artificially inflating the cost of said services to the end user.
Although Shaw management has made a statement denying that they will implement
UBB, the very denial itself reveals their intention to introduce it.
Such action is made economically possible through the independent and coordinated
abuse of substantial market power by a handful of large firms in the
telecommunications industry. Coordinated efforts include mutual artificial
constraining of supply of usage capacity, price fixing for overage charges, and
tacit collusion in the placement of usage caps on the various service packages
(also called limits or "thresholds" by incumbent PR departments).
The net result has been to force new entrants to the video-on-demand market to
significantly degrade the quality of their services (Netflix now has a special 'low
grade viewing' switch, designed specifically for Canadians due to the near
universal application of UBB among incumbent ISPs), to impair the ability of
new entrants to the video-on-demand market to compete effectively with
incumbents' own video and television offerings, and to artificially inflate the
cost for the end-user of video-on-demand services which compete with the
incumbents' own video-on-demand units.
Incumbent ISPs have further coordinated their efforts by teaming up and collectively
attempting to persuade the CRTC to impose additional levies on highly
competitive new entrants such as Netflix. Virtually all the incumbent cable
companies (most of which are also vertically integrated large incumbent ISPs)
cooperated in their efforts to pressure the telecom regulator into working on
the incumbents' behalf to impair Netflix' ability to compete in the Canadian
market. The effort was also meant to discourage new video on demand
entrants to the Canadian market by artificially impairing their ability to
compete in Canada as well.
Unless halted, the anti-competitive, collusive and cartel-like behaviour of firms in the
telecommunications industry threaten to thwart digital innovation in Canada,
and prevent consumers in Canada from having a quality of internet experience
on par with the rest of the developed world. It should be noted that Canada
stands virtually alone in the industrialized world in having usage caps and
punitive overuse charges being adopted by all incumbent firms in the industry.
Extensive research corroborating this is available from the OECD.
The imposition of the similar UBB schemes by all major ISPs cannot be considered
competitively neutral as it removes the ability of competing video firms to
provide competitive service. In many parts of Canada, there are only two
options for Internet service, a large vertically integrated incumbent cable
company, and a large vertically integrated incumbent telephone company.
The best way to achieve the stated goal of the Competition Bureau to "ensure that
Canadian businesses and consumers prosper in a competitive and innovative
marketplace" is to put an end to collusive behaviour among Canada's telecom
giants. Furthermore, the Bureau should take steps to ensure ISPs do not use
their pricing power to stifle the growth of video-on-demand competitors.
Furthermore, to ensure compliance with the Telecommunications Act, only cost
based pricing should be tolerated for regulated tariffs. This provide for more
stable and predictable pricing allowing service providers to design long term
business plans, while guaranteeing acceptable profit levels and low risk to
incumbents.
Therefore, the incumbent ISPs must be told to rescind usage based billing policies for
retail internet services, and abstain from such practices in the future.
It appears that any punitive fines may prove insignificant to the incumbents
Common misconceptions of bandwidth and data
transfer, and the myth of the bandwidth hog
To ensure the incentives and behaviours of firms in the telecommunications industry
can be properly monitored, a proper explanation of the transformation taking
place in the television and digital video on demand industries is needed to
establish the scope of this complaint.
Internet usage, that is, information transmitted over a network, is not a scarce
resource; information is infinite. Nobody is "hogging all the bandwidth" as is
often heard in the media. Bandwidth is simply not a consumable; it is a speed
of data transmission. Users pay for a certain speed of transmission, which they
cannot exceed due to technical limitations imposed by ISPs. The costs of
having a certain amount bandwidth available remain virtually the same
whether you use that available bandwidth to transmit 1000 gigabytes of
information, or you transmit only 1 gigabyte of information. The marginal cost
for the ISP of transmitting one more gigabyte over the internet been
conservatively estimated in a paper by Michael Geist, as 7 cents per GB, while
Reed Hastings has pegged the cost of transmitting one gigabyte as less than a
penny and falling.
To clarify some terms which have been confused by prominent people in the media,
senior officials in government and at the CRTC, and many consumers:
Bandwidth is not a consumable; it is a speed of data transmission. Data
transmission is the amount of data transmitted over a network. The term
"usage" is a weasel word which has confused many people, because it appears
to imply that data can be "used up"; it can't.
The internet connection inherently includes a usage cap determined by the speed of
the connection. Yet ISPs are trying to set an additional usage threshold above
which users are charged per gigabyte, with a markup somewhere between
5000% and 50,000% markup over marginal cost. They do this in concert, and
with remarkably similar pricing levels and structures far removed from the true
costs of providing the service.
The other problem with the UBB argument is the fact that unused portions of the
gigabyte allowance that consumers are paying for are not carried over to
future months. Thus consumers are penalized for using less than their cap
since they lose what they've paid for, and consumers are penalized for using
more than their exact allowance since they are charged exorbitant per
gigabyte fees beyond that monthly limit.
Some ISPs have attempted to use the excuse of network congestion to justify
introducing UBB. To this day, no proof has been offered to regulatory bodies
that congestion exists, and the Honourable Minister of Industry Tony Clement
as gone on record with a statement to that effect before a Parliamentary
committee in early 2011 (Standing Committee on Industry, Science and
Innovation). Any network congestion, if it exists, is a direct result of incumbent
ISPs overselling their services without sufficiently upgrading their infrastructure
to accommodate the influx of customers they have generated with their
advertising. service their customers have paid for in good faith.
The common cycle we see among incumbent ISPs in Canada is aggressive advertising
of high speeds, signing up of new customers, followed by ISPs receiving
complaints from new and existing customers about speeds not being as fast as
advertised, incumbents deflecting blame from themselves and attempting to
shift blame to the scapegoat of so-called "bandwidth hogs" -- who in fact are
their own customers who are simply making use of the bandwidth that they
have paid for. Incumbent ISPs then lower their internet use limits (or
"thresholds" in Shaw PRspeak) and raise charges for going above arbitrary
monthly limits on amount of data transferred over the network.
In some cases including April 2011, two or more incumbent ISPs announce massive
punitive charges and restrictions on supply of data transfer capacity at the
same time.
Even if we assume that network congestion does exist and is a significant problem, it is
only a problem at certain times of the day. UBB is akin to a car company
limiting all drivers to 60km of distance per month and then charging them $3
per km in excess of 60km per month, as a measure to reduce traffic congestion
during rush hour in the downtown cores of Canada's cities. The person driving
a 1000 km at 3 in the morning on a weekend contributes nothing to
congestion, but the person who drives home a much shorter distance from a
downtown office at 5pm on a Thursday certainly does contribute to traffic
congestion. Internet traffic is in a sense similar. The person who regularly
downloads 20GB at a time of HD films from itunes at 3am contributes virtually
nothing to congestion. However the person who gets home from school or
work at 5pm and immediately goes checks email and goes on Face book for a
few minutes contributes significantly to congestion. Yet under UBB the 3am
user is charged far more than the rush hour user, even though the 3am user
contributes nothing to congestion while the rush hour user is in fact
exacerbating the problem.
The nature of the emerging web-based video on
demand industry and the nature of the declining
legacy television broadcast industry
Traditional television subscriptions include a series of packages which are notoriously
disconnected from individual consumer interests. A consumer may wish only
to watch 3 channels: BBC, HBO, and the Movie Channel, but will be forced to
purchase hundreds of others at a cost typically exceeding $60 per month just
to access the channels s/he wants, even if the user has no interest in any of the
other channels. Virtually all channels are packed with commercial breaks;
often for every 2 minutes of content, there is 1 minute of advertisements. The
programs the user wishes to watch can typically be watched only at the time
scheduled by TV networks, or later on if the user has configured a PVR in
advance to record at a specific time. Traditional television subscriptions in
Canada typically involve lengthy contracts which punish users for switching
providers or cancelling their television service before all the years of their term
are complete.
Video on demand services such as Netflix offers a subscription service which allows
users to watch a vast range of television programming and film an unlimited
number of times, on demand at any time of day or night, without any
commercials, in HD, with no contract, for under $10 per month. It has proved
to be quite popular in the United States and has a rapidly growing subscriber
base in Canada. Around nearly 1 million households in Canada currently have
a Netflix subscription. Or 2.5 million people in Canada who may be accessing
Netflix in the home.
Canadian incumbent telecom companies have introduced their own digital video on
demand services, but not at a flat subscription price. Instead they have opted
to charge between 5 and 8 dollars per film, which is obviously uncompetitive in
the current marketplace.
The entry of Netflix to the Canadian market came after incumbents' launch of pricey
video on demand services. Furthermore, some companies including Bell were
in the midst of becoming yet more vertically integrated by adding Internet
Protocol Television to their offerings, and planning major acquisitions such as
CTV.
This appears to have been a concerted and direct attempt to squeeze Netflix, by
making Netflix' innovative, cost effective, flexible and desirable service more
costly for consumers.\
Further obviating the true intentions of the bandwidth caps (or limits or thresholds- all
the same thing) was the fact that the incumbents' own video on demand
services were exempt from the usage caps, even though a substantial amount
of bandwidth is dedicated to their own video services.
Incumbents' claims that Internet Protocol Television does not use the internet are
false. They do use a substantial portion of the finite amount of currently available
internet bandwidth. So the incumbents are charging UBB on Netflix video on demand
traffic, but no UBB on their own video on demand traffic. This is blatantly anti-
competitive and an affront to the spirit of the competitive market.
Incumbent ISPs are trying to delay the inevitable demise of their once-immensely-
profitable TV broadcasting business model by colluding to squeeze out
innovative new services like Netflix.
In areas of Canada where consumers have only two choices for essential internet
services -- a telco and a cableco -- the concerted squeezing of independent
video services through simultaneous identical UBB practices by the two
dominant players is in fact tacit collusion, and tantamount to cartel behaviour.
Netflix is not the only company ISPs are trying to squeeze by artificially inflating the
cost of using competing services. Amazon, Apple, Google and Hulu all have
successful video on demand businesses in the United States which are set to
expand into Canada.
Paragraph 13 of the November 29th submission1 by the Cable carriers in the 2010-803
consultation sums up the incumbents’ definition of UBB very well:
UBB is more about limiting growth of internet usage than about getting users to pay
for what they use. It should be called Behaviour Based Billing.
The incumbents are correct when they state that different UBB pricing will yield
different behavioural impact on usage. They are however wrong when they
insist that all ISPs must apply symmetrical level of ITMPs. The required ITMP
levels are dependent on target contention ratios which ISPs control through the
purchase of different amounts of capacity. Therefore, different ISPs need
different ITMPs because they do not all purchase the same amount of capacity
per customer. And different types of customers need different types of ITMPs.
It is therefore wrong to impose symmetrical ITMPs on all ISPs.
It should be noted that Bell Canada's throttling policy applies between 16:30 and 02:00
every day of the year across its entire network whether there is congestion or
not and cripples certain applications down to speeds which makes those
applications unusable 9.5 hours each day. Such throttling should be classified
as a behavioural ITMP as it conditions users to not use certain applications for a
greater part of the day. This is very different from a technical ITMP which
handles sporadic and localised congestion by dropping just enough of the lower
priority packets to deal with the temporary event.
Having 2 behavioural ITMPs causes some conflicts (double counting of traffic for
instance) and should have prompted the regulator to ponder the real intents of
Based on 2010-802 numbers, a user who consumes 80 gigabytes pays $30.00 of UBB
fees. A user who consumes 300 gigabytes pays $30.00. Since there is no
additional cost between 80 gigs and 300gigs, there is no incentive for users to
moderate their usage above 80 gigs. This defeats the stated purpose of
moderating the heavy downloaders.
With the pre-paid blocks, it gets worse. A user who buys 3 blocks of 40 gigs ($15.00)
ends up paying $45 in UBB fees if he consumes 300 gigabytes. So he is in fact
punished for purchasing pre-paid blocks.
A proper "user pays" model would see a linear usage curve with constant rates,
perhaps decreasing with higher levels due to a "volume discount". What Bell
Canada has introduced is not consistent with such a model.
2 The above 300 usage fees are approved, and Bell has provided notice they will
begin March 1 2011.
The proper definition of incumbents' UBB (cont)
Logic dictates that the punitive level of a UBB scheme should be proportional to the
odds of a user contributing to congestion problems.
The incumbents' UBB rates implement the exact opposite: more punitive pricing and
limits on users with the slower speeds which are far less likely to create
congestion. The 640kbs service is to have a microscopic 2gigabytes limit and
very high $2.50/gig charge while the 5mbps service, far more likely to cause
congestion due to higher speeds has a 60 gig limit with $1.50 per gig fee
thereafter.
All of the incumbents follow a similar UBB scheme, with higher charges for those on
cheaper plans, and lower charges for those on more expensive plans. This in
itself defeats any defence of UBB as a measure to reduce network traffic.
Paragraph 3 of Bell Canada’s introduction letter to the TN72933 filed on December 14th
2010 explains well the philosophy behind the counter-intuitive congestion
management pricing.
Therefore, the pricing is not set according to network congestion parameters but
rather by marketing parameters. The near duopoly situation results in each of
cable/telco incumbents following each other like a dog trying to catch its tail in
clear cases of tacit collusion.
3 TN7293 to replace TN7181 approved by 2010-802:
http://www.crtc.gc.ca/8740/eng/2010/b2_7293.htm
Incumbents pitch their UBB as a network management practice to control congestion
requiring exact pricing to ensure proportionate application of the ITMP. Then,
they admit their pricing is set by market conditions and are clearly not related
to any network management parameters.
Since incumbents exert almost total control over the market, their claim that they are
using market pricing needs to be rejected since incumbents set the market
prices.
The conversion to a UBB System (even if it were a fairly priced one) will greatly reduce
incentive to adopt new uses for the internet in Canada, while other nations will
progress at a more rapid pace.
By curbing end user usage, a punitive UBB regime allows incumbents to reduce the
pace of capacity investments.
To incumbents, UBB charges are a licence to print money since they are not tied to any
direct cash outlays. They have the flexibility to change them, offer promotions
or waive them to certain customers.
Competition issues (cont)
Pure play Internet Service Providers have not complained about rising internet usage
and welcome the challenge of serving customers with increased needs. And
because internet transit is a competitive arena, costs are driven down, so ISPs
have no problems purchasing extra capacity.
However, incumbents are quite different because they have monopoly in last mile and
have vested interests in legacy television distribution (even more now that
they also own broadcast networks) and know that the internet could do to their
TV business what iTunes did to brick and mortar music stores.
ITMPs such as UBB are a means to curb, delay or even prevent the adoption of these
new services, protecting incumbents' legacy TV distribution revenues.
Incumbents know that once an application has expanded beyond early
adopters, it is unstoppable. YouTube is a good example. So the goal is to nip
the TV competition in the bud before it is too late. To this end, the incumbents
are using coordinating their market power to ensure that no ISP gives the
market the choice between incumbents' walled garden legacy TV distribution
and innovative Internet-based entertainment.
From a competition point of view, the solution which supports a competitive market is
simple: investigate collusion, price fixing and cartel behaviour in the telecom
industry, and put a stop to it. This means preventing a group of entrenched
companies from all together enacting supply restrictions such as UBB and
behavioural throttling onto users of competing video services. Currently the
group of large incumbent ISPs are attempting to use their market power to
hinder certain new applications.
Competition in the retail ISP business also ensures there are open doors for new
applications and Internet-based services and that incumbents cannot prevent
the market from adopting services which compete against an incumbents' own.
Any measure which hinders Canadians' adoption of new internet application
will implicitly hinder the development of applications and services designed for
Canadians by Canadians.
Policy Issues
However, the UBB rates as presented by incumbents, being punitive in nature, share
very little with the true "user pays" paradigm.
Perhaps incumbents would change their rhetoric on UBB if they were faced with the
prospect of getting the reduced usage from ITMPs but not their revenues.
Policy Issues
Accuracy of usage
The 2010-255 CRTC decision touches briefly on the accuracy of the usage counting.
The CRTC accepted Bell Canada's statement that while there would be doubly
counted packets, this would represent a very small amount and the
Commission judged this to be appropriate. (paragraphs 41 to 44).
Independent measurements done in 2008 when the throttling was introduced showed
a sustained packet loss rate of between 20 and 25%, requiring that many
packets to be retransmitted. For packets bound to the internet, this means that
they are counted by the BAS before they are dropped by the DPI equipment.
In order to prevent speeds from rising, the DPI equipment must constantly drop
packets. This means that packet drops happens throughout a data transfer
using applications that the ISP has targeted for throttling. Since those
applications are used to transfer large files, the inaccuracy becomes
significant.