Daily Archives: November 30, 2010

Making Sense of Recent Internet Controversies: Level 3, Zoom, OpenDNS

The last weeks revealed three open Internet controversies. In these controversies, phone or cable companies abused their power over users’ access to the Internet to harm innovators and limit how consumers could use their Internet connections. These revelations come as the FCC and DOJ are reviewing the competitive impacts of the Comcast-NBC merger and as the FCC is (reportedly) about to issue a network neutrality rule. That is, these revelations come as the phone and cable companies are supposedly on their “best behavior.”

There are the three controversies, whose facts and implications I try to unpack below.

  1. A company called OpenDNS revealed to the Washington Post that Sprint Wireless is (and Verizon Wireless had been) blocking its services. In addition, industry white papers suggest that carriers should block services like OpenDNS based on the pretense of “security.”
  2. A company called Zoom Telephonics filed a complaint with the FCC against Comcast because, in short, Comcast is blocking consumers‘ ability to use Zoom modems to access the Internet.
  3. A huge company called Level 3 issued a press release accusing Comcast of imposing enormous charges on Level 3 when Level 3 tries to deliver Netflix’s streaming movies and other content to Comcast subscribers.

Telecom law students would recognize the first issue, OpenDNS, as a pure “net neutrality” issue. They would recognize the second issue, Zoom, as a “device attachment,” or “Carterfone” issue (in honor of an FCC decision by that name). But they would scratch their head in confusion and concern at the third issue, Level 3. This looks either like a common interconnection negotiation or a new kind of threat to the open Internet by a powerful cable company.

I will give the implications upfront. On substance, they reveal why the FCC needs to ensure any of its rules are not riddled with loopholes permitting threats to an open Internet. On procedure, they reveal why the FCC needs to adopt simple, inexpensive, rapid procedures for companies and consumers to file complaints on such issues, and wield a stick to deter these and future abuses.


OpenDNS is a direct competitor of all the phone and cable companies providing Internet service, as they provide DNS service as well. What is DNS service? It stands for Domain Name Service. Devices that are connected to the Internet have IP addresses, which are similar to phone numbers. So if you want to reach Facebook’s servers, they have an address. That address is a long numerical string. You will not be able to remember it.

To make addresses easy to remember, we have a way of translating those numbers into words, like “facebook.com” or “whitehouse.gov.” Some company, a “DNS service,” will translate the words you type into your browser into the right numerical address so you can reach Facebook.

Usually, your Internet Service Provider (say AT&T, Verizon, Comcast, or Sprint Wireless) will have a DNS service and it’ll be your default service. These companies sometimes make money when you mistype; you punch in facebk.com, which doesn’t match any number on the Internet, so your ISP sends you to a search page. On that search page are advertisements, for which the ISP gets paid. Voila, the DNS service makes money for the carriers.

You can choose an independent DNS service, like the one offered by OpenDNS or Google. Why would you do that?  Their DNS service might be faster. If you bounce around the web a lot, a faster addressing system would make pages load faster. Plus, these independent services may have innovative options for businesses and parents–for example, when your child punches in certain violent or indecent domains, the DNS could refuse to find such addresses.

I am not sure how long Sprint has been or is blocking OpenDNS and which other companies were doing so (Verizon is mentioned in the story). But Sprint’s incentive for doing so is obvious–Sprint could block a competitor that would take money out of its pocket. As Sprint Wireless was blocking a competing application, this would be a clear network neutrality issue.

Zoom Complaint

Zoom makes modems. What’s a modem? It’s that device at your home that connects your computer to the Internet. If you have Comcast as your Internet service provider, odds are that Comcast provided you a modem. But you could go to a store like Best Buy and buy your own. That modem could be made by Zoom or Motorola or others. While Zoom is one of the biggest sellers of modems this way, it is still just a small company trying to sell its products to willing customers.

Comcast has decided that it will block your ability to use a Zoom modem.

Why would it do that? If you’re using a Comcast modem, Comcast has more control over its features. Also, Comcast charges you a monthly rent for a modem, from which it makes a healthy profit.

The ability to attach any device to an Internet connection is the third principle of the FCC’s Internet Policy Statement. It’s also an old telecommunications principle beloved in FCC circles, usually referred to as “Carterfone.” The Carterfone case, decided in 1968, come from the era when AT&T was the national monopoly and your phone was hard-wired into the wall. You then rented your phone from AT&T for what amounted to thousands of dollars over the years. The FCC changed that in a series of decisions, and by the 1980s you could plug whatever you wanted into a standard phone jack, without AT&T’s permission, and without paying a rent. Since AT&T offered you a black phone … or another colored phone… the new options were extremely innovative. They included the answering machine, the fax machine, and the computer modem–and so making the Internet possible.

Comcast is violating the Carterfone principle, listed in the Internet Policy Statement. This reveals one threat to competition in devices, and therefore to innovation we’ve expected in the device market.

Level 3/Comcast

Level 3 just took on Netflix as a customer, and delivers Netflix’s movies and TV shows across its backbone to hand off to networks like AT&T and Comcast. As a result, subscribers of AT&T and Comcast Internet services would be able to watch the content they want to watch on Netflix. So Comcast made a “take it or leave it” threat to Level 3, forcing Level 3 to pay Comcast a lot more money to deliver what would be more traffic. These costs would likely be passed onto Netflix, which would weaken Netflix–a competitor to Comcast’s cable TV offerings.

This is the one confusing people right now. I have been on dozens of email strings, on several listservs, about the significance of this issue. I have seen investor reports, spoken to journalists, lawyers, researchers. It reminds me of when Apple’s iPhone refused to accept Skype into the App Store–people weren’t sure what to make of it initially. With Apple and Skype, that seemed like an anticompetitive action. But people didn’t traditionally worry much about choices made by device-makers like Apple, because the device market is generally competitive. Rather, people had worried about the phone or cable companies exercising their last-mile power over Internet users in a way to affect the potentially competitive markets of devices (see Zoom) and applications (see OpenDNS, among others). In the end, it turned out, that AT&T itself had been involved in the Apple decision to exclude Skype–so the AT&T threat was present.

The confusion with the Level 3 situation is that people usually think of the backbone market as competitive, or at least competitive enough not to warrant much regulation. What’s the backbone market? It’s kind of like the long distance lines for the Internet, connecting all the local networks. Because they connect whole cities, and don’t require building a line to every single home in a neighborhood, the cost structure of the backbone is such that you could conceivably have several competitors. (If you remember, we had companies like MCI and Sprint, in the 1980s, competing in long distance, though not in local calls, because of the costs of building the long networks versus the local networks.)

The FCC has not wanted to regulate the backbone market. It issued an excellent, kind of “telecom-famous”paper about this issue in 2000, called the Digital Handshake.  The main arguments against regulation were: (1) there was competition in the backbone market, so regulation was unneeded, and (2) it ain’t broke, don’t fix it. A lot has changed in the past 10 years. The backbone market has become more concentrated, as has the ISP market. And technologies have advanced, permitting certain discriminations in the backbone transmissions that were not possible 10 years ago. An independent problem is: we know very little about the market, as the FCC gathers little information and the agreements are generally confidential.

Level 3 is a backbone. It engages in commercial deals with other backbones and end-user ISPs to exchange and deliver traffic. In Comcast’s version of the story, Level 3 is just trying to get a better deal than backbones usually get for delivering traffic. Ordinarily, if this were true, Comcast would have a point.

What complicates this story is that Comcast is not just a backbone in a competitive backbone market; it is also an ISP that controls access to the Internet in 16 million households in an uncompetitive market usually controlled by one local cable monopoly and one local phone monopoly. Moreover, cable companies dominate the market for local wireline Internet access. Phone companies cannot compete well because cable technology enables faster Internet access. Plus, not all phone companies can offer TV and Internet in a “bundle” like cable companies can. So Comcast has what economists call a termination-access monopoly over 16 million homes, meaning if Level 3’s customers want to reach those 16 million people, Level 3 has no choice but to deal with Comcast. Comcast has essentially a monopoly over those people (even if a temporary one).

Comcast’s actions reflect the usual concern in telecom regulation: a company leveraging its market power in the last-mile network (traditionally the least competitive part, and the one least susceptible to competition). It is leveraging that power into the parts of the network that are potentially competitive: the backbone market and the application market. This is exactly the network neutrality concern and the Carterfone concern.

What application will be affected? Netflix. And so will any online TV company that competes with Comcast cable offerings, its video on-demand offerings, and its online on-demand offerings.

So–is this just a backbone dispute or a threat to the open Internet by a last-mile provider?

It looks like a threat to me, but the facts are still emerging.

What to Do?

The implication of these violations is clear: we can’t put our head in the sand and hope that the Internet remains the amazing general-purpose engine of innovation, speech, and competition it has been traditionally been.

The FCC–our telecommunications agency–needs to ensure it has the jurisdiction and substantive rules to address new and dangerous threats to users’ and upstarts’ ability to reach consumers through the telecommunications networks. Very simply, the rules should ensure that the FCC can reach actions by phone and cable companies that threaten Internet freedom, even if those actions are laundered through contracts with phone-device makers like Apple or backbone contracts. The FCC need not predict these actions in advance; phone and cable companies will be creative in their attempts to avoid any rules that exempt, say, undefined classes of “managed services.”

Beyond that, upstarts and consumers need a place to go where they can file complaints, seek redress, and get legal clarity. The FCC needs to ensure that its rules for addressing complaints are cost-effective, quick, neutral, and relatively simple. Otherwise, with labyrinthine FCC procedures, the billion-dollar  phone and cable companies will be able to use their political power and vast wealth to wear down and defeat any upstart. Unless the FCC is willing to act, and impose severe penalties, it’ll be cheaper for phone and cable companies to block competitors than to invest in innovation.

The FCC has an opportunity to address these issues in an upcoming net neutrality rule. It should take that opportunity, as the carriers may soon not be on their “best” behavior.

Models for the Internet’s Future: Obama-Open or Julius-Closed

Apparently before the year is out, on Dec. 21, the Federal Communications Commission will issue rules to help shape the future of  the Internet. In fact, the FCC Chairman may be circulating those rules to fellow commissioners on Wednesday, tomorrow. These rules will decide how much control AT&T, Verizon, and Comcast will have over the websites you can visit and the online software you can use. This rule will impact the future of businesses, political actors, and people who now rely on an uncontrolled, open Internet.

There are at least two competing regulatory models for the FCC to adopt.

One is a model being pushed by AT&T and Verizon –also known as key opponents of network neutrality. This model derives from an attempted compromise offer from Congressman Henry Waxman to congressional Republicans. The proposal, never introduced, failed to gain Republican support–but the FCC Chairman does not need congressional Republican support on a Commission that is majority Democrat. AT&T has been meeting repeatedly with top FCC staff to push this option, after spending five years and hundreds of millions in lobbying fees to oppose real network neutrality protections.

The other model comes from an agency controlled by President Obama, called the NTIA (or National Telecommunications and Information Administration), which is less well known than the FCC, an “independent” agency not under the president’s direct control. Early in this administration, the Obama NTIA implemented tough rules to ensure Internet freedom on all private Internet networks under that agency’s jurisdiction–those networks were those receiving even a penny of stimulus money under the NTIA stimulus program.

The FCC could choose rules favored by AT&T and Verizon … or those adopted by the Obama’s own agency, in keeping with Obama’s campaign promises and the public declarations by the FCC itself.

So, for the benefit of all of you who have not spent much of the last five years thinking about (and lawyering on) network neutrality, I figured I’d include a primer on the two potential models–the Obama NTIA model and the other, weaker model, urged by AT&T.

Network Neutrality Models–Generally

Any “network neutrality” rule should be designed to forbid phone or cable companies from controlling the Internet. Evidence and economic theory suggests that control of the Internet by the phone and cable companies would lead to blocking of competing technologies (as in the Madison River case), blocking of innovative technologies that may not even compete with the phone/cable cartel (according to Comcast itself, the Comcast/BitTorrent case would be an example), and increased spying on Internet users. It would have major effects on speech by raising the costs of speech for campaigns, individuals, and organizations, as well as the blocking of “controversial” speech by religious groups or abortion-rights activists.

There are many models (and factors) for implementing a network neutrality rule, but we will discuss two. For completeness, some of the other models include: the Republican FCC’s 4-principle Policy Statement, the Republican FCC’s “strict scrutiny” elaboration in the Comcast, the principles found in the Democratic dissents-and-concurrences to Republican FCC orders, the Obama campaign promises and presidential pledges, the NTIA rule, the somewhat unspecified Genachowski NPRM proposal, the Verizon-Google pact, the Henry Waxman proposal, the French rule, the Canadian rule, and others.

We’re only discussing two here: the NTIA rule (because it represents the Obama position) and the “Waxman proposal” (because AT&T is pushing this proposal hard at the FCC).

The key factor for such models is whether they actually preserve an open Internet or whether they have gaping loopholes. The main loopholes would be

  • exceptions for wireless Internet, an exception unjustifiable based on consumer preferences or technology, especially while the FCC is betting on wireless technologies to expand broadband access,
  • defining unlawful discrimination in a narrow way that would let AT&T and other carriers charge companies for priority treatment, even though a ban on paid priority is generally considered a basic principle for any meaningful net neutrality rule,
  • defining unlawful discrimination to forbid targeting only those technologies that already compete with services offered by phone and cable carriers; this would allow carriers to target non-profits, political speech, and technologies too new to compete with the carriers’ services;
  • a blanket exception for still undefined “private Internet” services,
  • and an exception for “network management” that swallows the rule.

In addition, a network neutrality rule could result in mere “slaps on the wrist” or involve such expensive and difficult litigation procedures that no small company or consumer could ever bring a case. Finally, it could rest on clearly flawed jurisdictional grounds.

Model I:

The first model is even worse than the widely-criticized Verizon-Google pact and a far cry from the Obama/NTIA rule. This model is one that:

  • fails to protect wireless Internet access. This proposal would forbid carriers from blocking technologies only if a broadband provider already had an interest in a competing application. As a result, AT&T and Verizon would be able to block anything truly new or innovative–that is, anything the carrier had not already deployed or even yet imagined. Essentially, it would ensure that the most innovative applications get the least protection.
  • defines nondiscrimination vaguely, and is silent on whether paid prioritization would be banned. Even the Verizon-Google proposal had a ban on paid priority.
  • expired in two years anyway.

So, in the end, Model I is not even a model for net neutrality; it’s make-believe. It’s a model of a Congressman trying valiantly to help a Chairman and reach a compromise with other elected legislators in Congress. It is not a model for “consensus” with an unelected industry when a regulatory agency has the necessary votes after many years of study and an enormous record.

Model II: Obama-Open, Obama-Tough

The second model is better for America and all of us relying on an open Internet for commerce and speech.

President Obama is a big supporter of keeping the Internet open. During his presidential campaign, he pledged his support to net neutrality repeatedly (on MTV, at Google) and made net neutrality a centerpiece of his technology agenda. Net neutrality was, in fact, central to his argument to Silicon Valley against his primary opponent, Hilary Clinton. As a senator, he co-sponsored the lead network neutrality bill.

Obama’s campaign promises are explicit on several details where the carriers have tried to find loopholes: Obama would forbid carriers from imposing any “toll charges” (also called “access charges” or “paid prioritization”); he hints at no exception for wireless Internet access; and he would forbid online discrimination regardless of anticompetitive effect, protecting non-profit and individual speakers.

As president, Obama pledged support for net neutrality in rolling out his cybersecurity agenda and in answering the “most commonly asked question” about the economy (which was on net neutrality) in a Youtube interview. He was unequivocal, and opposed paid prioritization there, despite “pushback” from the biggest carriers. White House officials told Time that a net neutrality rule must cover wireless technologies.

Beyond this, and more importantly, President Obama actually imposed a network neutrality rule. He implemented a real rule through the stimulus bill. The two executive agencies that decided which companies received stimulus money for broadband networks–the NTIA and the Rural Utilities Service–imposed strict network neutrality rules on those who received this stimulus money.

The NTIA is also the agency that speaks directly for President Obama; it is known as his “voice” on telecommunications. For example, when the administration files its thoughts on telecommunications in any proceeding of an independent agency (even the FCC’s proceedings), the NTIA that files those thoughts on the administration’s behalf.

The NTIA neutrality conditions apply to both wireline and wireless networks, include a nondiscrimination mandate, and have tight language on any exception for network management. For example, on nondiscrimination: applicants could “not favor any lawful Internet applications and content over others.” And management of networks appears limited to the carrier addressing harmful attacks and managing congestion in a way that does not treat applications and content differently.

There is no reason why strict net neutrality rules should be limited only to companies accepting stimulus funding. Net neutrality advocates, including President Obama (and Senator Obama in 2007 and 2008), have never limited their proposals to stimulus networks. In addition, phone and cable companies receive huge subsidies from the government, through tax write-offs for accelerated depreciation and the FCC-administered universal service fund. Both cable and phone networks were built and maintained for decades under monopoly protection, with government-guaranteed returns, an advantage upstarts lack.

Perhaps the most important thing to know about Model II is this: the FCC consulted the NTIA and RUS on the rule. The stimulus bill  (Section 6001(j) of the Act) required “coordination” with the FCC.

I am not sure how this coordination went, but I highly doubt the FCC tried to water down President Obama’s rules then, for those networks. It shouldn’t go for watered down rules now, when the Internet’s future depends on a solid rule.

Cross-posting at Balkinzation, Huffington Post, Stanford CIS blog, and The Faster Times.

“Holy Grail” of Networking Found?

Ben Rooney writing for the WSJ’s Digits Blog reports on a breakthrough made by the Irish company Intune Networks – the optical burst packet switch, which could improve fiber efficiency upwards of 80%.

Did Stuxnet Disable Iranian Centrifuges?

Word broke Monday that Iranian President Mahmoud Ahmadinejad had admitted the Stuxnet worm had disabled some of  his country’s uranium centrifuges.  But Computer World reports on some skepticism of the report, noting that some researchers are waiting for more proof (however difficult it might be to acquire it) before concluding that the worm did in hamper the nuclear program.

Did Google Get Hacked Because a Chinese Official Didn’t Like His Search Results?

The Chinese government has been running a hacking ring that launched a hack attack on Google earlier this year, according to a diplomatic cable leaked by WikiLeaks yesterday.

Web Delivery Firm Says Comcast Charging Unfair Fee For Data

Level 3 Communications Inc., an Internet backbone company that supports Netflix Inc.’s increasingly popular movie streaming service, complained Monday that cable giant Comcast Corp. is charging it an unfair fee for the right to send data to its subscribers.