December 18, 2014 4:31 PM
[Note: This is the first post, in a series, in which we'll look at the Netflix/Cogent/Comcast congestion episode from earlier this year. The focus will be on understanding this event from a different perspective than most of us may have thought about it before. This series looks further into the question I raised in my last post which is: are existing laws--adequately enforced--sufficient to protect consumers? For purposes of readability, the full citation of relevant facts has been placed at the end of this post.]
In my last post
, I started to look at whether the protracted congestion--and associated consumer service disruptions--caused by the recent Netflix/Cogent/Comcast interconnection dispute indicated that the traditional voluntary agreement structure of the Internet was broken, or whether existing laws might not be enough to prevent protracted consumer disruption. In a recent article
, Prof. Susan Crawford, an advocate for all things Net Neutrality, also highlights the frustrations of customers caught in the middle of the Netflix/Cogent/Comcast dispute,
No federal, state, or local government exercises any oversight over this handful of interconnection points. No Better Business Bureau watches over how your requests for data are being treated.
Prof. Crawford is right to question why consumers got stuck with the short end of this wholesale dispute. But, I disagree with Prof. Crawford's assumption that new laws--specific to the issue Netflix and Cogent blame for the protracted congestion--are needed.
As noted in the last post
, more specific contracts, and quicker enforcement by wholesale partners are one way to prevent extended periods of consumer frustration. Similarly, there are also existing laws designed to protect consumers from intentional, or knowing, actions of third parties that prevent consumers from receiving services they believe they are purchasing.
The FTC has expressed
concern that it will lose jurisdiction over Net Neutrality-related matters if the FCC decides to reclassify broadband as a Title II service (the FTC Act specifically exempts "common carriers"). Even though I was aware the FTC had asserted
jurisdiction to handle Net Neutrality complaints, I didn't really think about the how important the FTC could be. . . until I started looking really closely at the Netflix/Cogent/Comcast congestion episode from earlier this year. "It's About the Principle"
When Netflix filed its Petition
to Deny the Comcast-TWC merger in August, I was interested in learning the circumstances that led to Netflix's direct interconnection agreement with Comcast. I expected to see a pretty basic recitation of how Comcast kept unreasonably increasing interconnection prices, thereby forcing Netflix into lower quality interconnection arrangements.
Instead, though, the brief spends a lot of time establishing that Netflix's principle
--of not paying the ISP any portion of the costs of delivering its content to its customers--was the exclusive factor it relied upon in choosing vendors to deliver its customers' traffic. The absence of any comparison of various input prices/vendor alternatives available to Netflix seemed odd. The notion that Netflix was defending a not-purely-economic principle seemed odder still.
The antitrust analytical framework (which was the ostensible basis for
Netflix's opposition to the Comcast-TWC merger) recognizes economic
efficiency, and not any unique firm-specific view of how an industry
should work. Yet, Netflix has never indicated that its decisions were
based on immediate cost/price effects. It has even clarified that the
costs being imposed by the ISPs are not significant, nor has it raised
prices for customers served by the offending ISPs. See, e.g.
, this blog
. Level 3 made the same argument in 2010--that it's not the cost, it's the principle. See this Ars Technica article
Prof. Susan Crawford, in the article
mentioned above, also observes:
FCC will find that the money amounts involved in these deals are low at
the moment. It's the naked threat posed to the future that is the
problem. . . .
The "naked threat posed to the future"
may or may not be cause for concern, but--if this threat does not limit
the immediate ability of a firm to deliver service--can a firm's reaction
to such a threat excuse its performance under its customer contracts?
possible that, for many months at least, Comcast customers (and those
of certain other ISPs) were paying for service that Netflix knew to be
substandard. If Netflix failed to take any action to provide the grade
of service for which its customers were paying, or to let prospective
customers know they would be receiving degraded service for an
indeterminate period, then it's possible that enforcement of existing
laws might prevent future consumer abuses. The Federal Trade Commission Act
5 of the FTC Act prohibits "unfair or deceptive acts or practices" that
affect commerce. An act or practice may be found to be unfair where it
"causes or is likely to cause substantial injury to consumers which is
not reasonably avoidable by consumers themselves and not outweighed by
countervailing benefits to consumers or to competition." See, FTC "Unfairness Statement
The FTC is likely to find an act or practice to be deceptive if "there
is a representation, omission or practice that is likely to mislead the
consumer acting reasonably in the circumstances, to the consumer's
detriment." FTC Deception Statement
standards for unfairness and deception are independent of each other.
While a specific act or practice may be both unfair and deceptive, the
FTC may find a violation of Section 5 if the act or practice is either
unfair or deceptive.
In the next post, we'll look at what
happened and whether Netflix's 2013 change in the way it delivered
content to the country's largest ISPs seemed reasonably calculated to
efficiently deliver content to customers, or, if not, whether it seemed
designed to promote another goal. ***Relevant Facts***
is a brief recitation of the relevant facts for purposes of our
discussion. Unless otherwise cited, the facts are taken from the
Declaration of Ken Florance , Netflix's Vice
President of Content Delivery, submitted in support of Netflix's
Petition to Deny
the Comcast-TWC Merger (FCC Docket No. 14-57).
most of the history of Netflix's streaming video delivery service,
Netflix believes that Comcast has required Netflix's third party vendors
to pay an additional fee to cover some (or all) of the cost of
Netflix-specific capacity augmentation at interconnection points.
Netflix describes 3 instances between 2009 and 2010 where it believes
CDNs needed to purchase additional capacity to alleviate congestion
issues. Florance Declaration ¶¶ 29-41.
Netflix acknowledges that
the volume of its traffic does increase demand for ISP-bound capacity at
its vendors' points of interconnection with Comcast. Moreover, these
costs are incremental and specific to the particular point of exchange
between Netflix's Internet transit vendor and the ISP. Florance
Declaration at ¶ 46.
When its traffic was carried on third
party CDN networks, Netflix was aware of the costs being incurred on its
behalf, but "in the short term Netflix was insulated from a sudden
price increase." Florance Declaration at ¶ 39. While Netflix was using
CDNs, its performance over cable systems seemed uniformly better
even on the most advanced telco systems.
its service was good using 3rd party CDNs, Netflix explains that,
"[a]fter the Akamai, Limelight, and Level 3 CDN congestion episodes
[2009-2010], Netflix began transitioning its traffic from CDNs (all of
whom, we believed, were paying Comcast's new terminating access fee) to
transit providers in our continued effort to avoid terminating access
fees." Florance Declaration at ¶ 40. (dates in brackets added). Thus,
in February, 2012, Netflix signed an agreement with Cogent for Internet
transit service. Cogent began transitioning traffic to Netflix in
August 2012. Florance Declaration ¶ 41.
Based on customer
complaints about service quality, Netflix's service deteriorated
immediately upon switching to Cogent transit and progressively
deteriorated over the next year. Florance, at ¶ 51. However, beginning
in October 2013, Netflix reports a very high level of customer
dissatisfaction and cancellations, due to "Netflix's inability
anything to change the situation." Florance ¶ 52 (emphasis added).
Continue reading The Netflix/Comcast Dispute: Interconnection "Principles" vs. Consumer Rights?
February 20, 2014 2:38 PM
I was hoping that the Chairman would resist the temptation to open a new net neutrality rulemaking after last month's D.C. Circuit decision
overturning most of the FCC's 2010 Open Internet Rules--but he didn't
. The Open Internet Rules
were unnecessary when they were adopted, but more importantly, they were a waste of a significant part of Chairman Genachowski's tenure.
The FCC already has to deal with the devilishly-complex, two stage spectrum incentive auction/reverse-auction
. Similarly, the Commission is already lagging the market in considering what regulatory modifications may be appropriate for consumers in an "all-IP
" world. The Commission is also likely to spend considerable time and resources reviewing the proposed Comcast-Time Warner Cable merger
If the FCC really thought about whether net neutrality rules are even useful, much less necessary, it would quickly conclude that the "terminating access" theory (underpinning the arguments of net neutrality advocates) should probably be left back in the MFJ--where it came from. Making rules to thwart hypothetical problems is--at best--a distracting waste of time. But, when the putative rules will affect services that don't presently exist, the danger of real harm becomes much more likely.
Why are net neutrality rules unnecessary, and even potentially harmful to the productivity of the Internet?
. The Internet did not become "open" by accident. With the possible exception of P2P traffic
, almost every form of traffic carried on the Internet is delivered pursuant to a contract that specifies a guaranteed level of carriage.
Think about it. Websites use a type of "ISP" called a web host, usually a data center connected to multiple backbone providers. If the website will be serving up a lot of traffic to its customers, the website will frequently use a "content delivery network" (CDN
) to ensure that it's traffic is delivered in the fastest manner possible. Many large Internet backbones (like AT&T
, Level 3
, and Verizon
) also provide CDN services.
In fact, most large ISPs also provide significant services (hosting, Internet backbone, CDN, and "cloud services" for large enterprise customers) which depend on the assumption that the consumer's ISP will carry the tendered traffic in a non-discriminatory manner. Any act of website-specific discrimination by an ISP could easily be detected, will likely put that ISP in violation of its peering contracts, and will invite an avalanche of against the offending ISP.
The bottom line is that--even if an ISP had the ability and the short-term economic incentive to discriminate against another carrier's Internet traffic--the consequences of the discrimination are neither predictable nor quantifiable. If economics is to be believed, the ISPs are extracting all the revenue the market will bear--further price increases would only reduce profits.
2) The Success of the Internet of Things Might Depend On Discrimination
. Many were surprised when Google paid $3.2 billion last month to acquire
smart thermostat company Nest Labs, but this type of service is a critical part of the "Internet of Things." An energy utility could realize significant benefits by receiving real-time consumption data from households. If the energy company could anticipate, and alleviate, peak period demand spikes--perhaps by remotely adjusting appliance demand for those customers willing to participate--the company could reduce its costs for expensive peak capacity.
The value of telemetry depends on the utility getting timely information from a significant number of households, but no one wants to pay for their air conditioner's bandwidth. However, your appliances--and the energy provider--can tolerate data service that is too slow or jittery to support a latency-sensitive application (like VoIP). So, if the bandwidth for appliances was cheap enough, there are probably many "win-win" applications that someone other than the Internet subscriber would be willing to subsidize.
Likewise, those big software updates
and gaming patches
could be delivered in a way that is cheaper for both the ISP and the consumer, if the provider or the consumer were offered a time of day/de-prioritization option. Discrimination isn't bad if it's just an option. But it's an option that "rules" tend to discourage, if not foreclose.
3) The Consumer Can Always Evade Discrimination
. According to Sandvine's data for both the first
half of 2013, one of the most significant Internet traffic trends over the last year (for fixed and mobile North American networks) has been the growth of "tunneling" traffic. Tunneling
refers to customers using VPNs
to obscure their content when accessing the Internet.
The VPN encrypts the customer's traffic and routes it to a server which assigns a random, or sometimes shared, IP address. Thus, all of the customer's Internet traffic originates/terminates through an "anonymous" IP address at a server remote from the customer's home computer. To the ISP, it simply looks like the customer is sending and receiving a lot of non-destination-specific traffic to a smaller number of IP addresses.
Prior to concerns over privacy, tunneling was most frequently used by consumers for online banking, and employees working from home to access their company's networks. Whether tunneling will protect your information from the government is unclear, but the existence (and forecasted) growth in tunneling traffic will serve to protect you from hypothetical fears of discrimination by your ISP--even if P2P is your thing. So Why Are New Rules Needed?
If the content providers are protected by contracts, consumers can protect their traffic from ISP inspection through encrypted VPN tunnels, and new consumer benefits can be realized from efficient, permissive discrimination, then it wouldn't seem that there's a whole lot to be gained from a proceeding to add to the remaining disclosure rule. Given the immense opportunity costs of diluting agency focus at this moment, let's hope the chimerical fears of a few do not capture the public's scarce regulatory resources. The FCC can best protect the Internet by focusing on the IP transition and bringing more wireless spectrum to market.
October 11, 2013 12:58 PM
Earlier this week, the Internet Innovation Alliance
released a study
on old networks, new networks, wireless networks, red networks, blue networks; who's using them, and how much they cost. Coming in at 45 pages, including numerous pictures (each worth ~1000 words), the IIA had more to say than Sen. Ted Cruz on bath salts. Luckily, you've got me to unpack this baby for you.
In short, the IIA report dramatically shows what every FCC commenter has ever said in support of their comments: and that is that if FCC adopts the regulatory (vs. de-regulatory) policies advocated, then these policies will promote investment above and beyond the level necessary to deliver the regulated service. To be clear--FCC regulations
(they have to be affirmative burdens on regulated firms) promote investment.
Let's take a closer look. According to IIA, only 5% of households depend on "POTS" (plain old telephone service), and the switched telephone network handles only about 1% of the voice traffic handled by IP networks (wireline + wireless + cable + CLEC). Yet, and here's the kicker, from 2006 through 2011 more than half
of incumbent LEC investment was used to support the POTS network in order to comply with . . . wait for it . . . FCC regulations!
This investment--which was clearly not necessary to deliver voice service to consumers--amounted to over $13.5 billion dollars/year. This is investment that can only be attributed to successful regulation. I'd say former Chairman Julius Genachowski has something else to crow about . . . as if he needed another feather in his cap.
The impact of this report cannot be understated. The Commission has scarcely seen such vindication of its efforts. While the IIA tries to shy away from its pro-regulatory conclusions by saying that but for the FCC's legacy rules, more resources would have been diverted to providing advanced IP services that consumers want to use, this is idle speculation from self-interested parties. The IIA knows full well that maintenance of this "museum network" is critical to our country's economic recovery.
Interestingly, the IIA study confirms what the left has been saying for years: cable needs to be regulated. Why? Well regulation certainly won't make them better companies--apathy is in their DNA--but regulation will make them pay their fair share. Take, for example, Comcast's customer service rating
: a resounding "disappointing", but only a step away from "terrible." Verizon comes in almost even
(the result you would expect in a competitive market).
The only difference between the two companies? Verizon is contributing to the economy, and Comcast is getting a free ride--regulation-wise. If Comcast were regulated, would they all of a sudden improve by providing "mediocre" or "somewhat acceptable" customer service? No! Of course, not--they would in all likelihood remain competitive with Verizon, but they would be contributing to US economic development.
Perhaps the party that comes off looking the worst out of this study is US Telecom
. US Telecom, in case you don't know, is the trade association of the incumbent LECs, and US Telecom has repeatedly fought the Commission's efforts to preserve these pro-recovery, pro-investment regulations.
US Telecom filed an extensive petition
last year explaining why a ton of legacy regulations no longer serve any useful purpose. In fact, US Telecom has another petition
pending with the Commission right now, seeking to remove some of these regulations by having its members declared "non-dominant" in the provision of wireline telecommunications services.
In its arguments, US Telecom conveniently fails to list "investment" [for the sake of regulation] as a useful purpose of regulation. This omission, of course, makes the US Telecom requests look deceptively reasonable
Hopefully, the Commission will see through this shameless ploy and do what's right for the economy. Frankly, it matters little whether 5% or .0005% of consumers use incumbent LEC wireline services: dominant is dominant, and dominant = regulations, dammit. You know who "gets it?" The competitors of the US Telecom members, that's who.
Because of the government shutdown (costing countless jobs from regulations that are simply not being adopted), I couldn't check the FCC website to see who else had the foresight to oppose the US Telecom petition, but I did manage to find these opposition comments from CompTel here
. Thank goodness someone cares about preserving the investment
incentives of their competitors.
February 17, 2010 5:59 PM
With sincere apologies to the members of the Google Nation, let me be clear about my last post
. I was not "hating on" Google. My only point was to try to mollify some of the "irrational exuberance" that emerged on the Net (and in the press) as a result of Google's understated "announcement" of its plans for a broadband experiment. For those that didn't read my last post, one week ago (Wednesday, February 10th), Google stated on their corporate blog that they would like to build a fiber network to deliver 1 Gigabit speeds to anywhere from 50,000 to 500,000 homes. Most ensuing stories on the Net and in the press reported on/reacted to this announcement as if the project was already under construction.
For those who want to believe in the existence of a "Google-Claus", I strongly recommend the dose of reality that you can get from reading Harold Feld's post from yesterday
, where he does an excellent job of providing a detailed account about Google's success through the years of "bluffing" and "slow-playing" regulators and network operators in order to get network operators and their end-users to front the cap-ex to support the transmission speeds that will enable Google to offer more services with which to economically advance their business. There is little reason to believe that this announced "experiment" will bring Google any closer to being a broadband ISP than any of their previous rhetoric.
On the other hand, Google has been quite straightforward about their business plan,
which is to create applications that allow them to capture more and more customer information that they then "monetize" through (essentially) resale to advertisers. Therefore, I come not to bury Google, but to praise them . . . for their honesty in dealing with users of all their services, including Google "Buzz" (which coincidentally was really launched on the same day that their broadband network plans were announced). In a reaction that is surpassing strange, the outrage on the Net and in the "blogosphere" over Google Buzz is comparable to the enthusiasm surrounding Google's 1 Gig "broadband network."
But why do I say the outrage about the Google "Buzz" product is as perplexing as the enthusiasm over the non-existant, broadband network? Well, it's simple. Google has never been in the business of being a telecom network operator. In fact, if Google has read the newspapers over the last 10 years--and it's clear they have--we can assume Google knows that entering the retail broadband Internet access market (even at efficient scale) is very often a good way to make a small fortune (out of their current large fortune). To the contrary, though, Google is in the business of obtaining and selling Internet user information.
On this point, Google could not have been more clear with users of its products. Only two months ago, Google's CEO, Eric Schmidt, told Americans--on a national cable network--in a statement that was widely repeated
, something to the effect that if consumers don't want people to know what they're doing online, then they shouldn't be doing it [at least not using Google services] in the first place
. To underline a point, shortly thereafter, one of the founders of Google's major search partner--Mozilla Firefox--encouraged users to switch to Microsoft Bing for privacy reasons
In short, if consumers decided to continue to avail themselves of Google's "free" services (like Gmail or Google Search), even after Google's December clarification that consumer privacy concerns take a back seat to Google's policy of using consumer information generated by use of its products for its commercial purposes, then it's a little difficult to understand all the "outrage" surrounding Google's Buzz product. When one considers that many of these same critics are also arguing for rules to keep the Internet "open", the complaints are even more difficult to indulge. Do these outraged privacy watchdogs really want an "open" Internet, or just an extension of the "Nanny-state" that relieves them of any personal responsibility with respect to how they use the Internet?
On this controversy, Google is in the right. They've given consumers enough information to make up their own minds. If consumers choose not to use this information, then what is the point of an "open" Internet? What is Google's incentive to continue to innovate and provide "free" services to those customers that have nothing to hide, and are happy to trade information for applications?
If we regulate Google's online behavior, next thing you know, we're regulating the ability of legitimate
Nigerian businessmen to use the Internet to raise capital--just to get at the few fraudsters that abuse the gullibility of some Internet users. But how does this "outrage" do anything to promote commerce, jobs, innovation and openness? It doesn't, and it's about time for the "Internet police" to dial back the schadenfreude
, and lay off the last guardian of the open Internet.
January 13, 2010 9:56 PM
Last Friday, the 8th, I did a post
on the reports about the Comcast v. FCC
oral argument that was held before a skeptical D.C. Circuit that morning. The point of my post was that the Net Neutrality NPRM
(comments are due tomorrow!) might be a "rainout", because most reports suggested the court was less than encouraging about the Commission having authority to enforce its Broadband Policy Statement
, based on the two main statutory provisions the Commission relied upon in both defending the Comcast decision
, and supporting the current NPRM and proposed rules. If the court did vacate the Commission's authority to enforce its Policy Statement, or any similar Title I rules, then--my post noted--the Commission would have to start again with new rules based on different statutory authority.
I also noted that, when asked if it would rather lose on "narrow" (did Comcast have adequate notice?) or "broad" (do the statues the Commission relies upon, really provide the authority to regulate specific Internet practices?) grounds, the FCC said it would prefer a narrowly-written loss. I failed to note that Comcast agreed. While I figured a "broad" loss for the FCC would be bad for cable, it seemed kind of speculative and I really didn't want to get into it.
On Monday (the 11th), though, Harold Feld waded into the topic with an excellent post, entitled Does Comcast Fear To Win Too Much?
In this post, Harold confirms Comcast's fears by citing a back-pedaling post
that appeared on Comcast's policy blog Monday. The post was an excessive "clarification" of Comcast's "true position" that the FCC does
have the authority to regulate Internet practices under Title I. Now a Shakespeare aficionado might observe, "[Comcast] doth protest too much, methinks.
" But, really, who cares what Comcast thinks? The court's interpretation of the scope of the Commission's authority is going to come out sooner or later, anyway.
Continue reading Comcast Isn't the Only One Afraid of a Big Win