I keep seeing a steady stream of articles that basically go through the following analysis:
1. ‘Net Neutrality is about treating all bits equally.’
2. ‘The Internet has never done this. Particularly we have this thing called ‘content delivery networks’ or ‘CDNs,’ that have been a vital part of moving content around the Internet for over 15 years. CDNs are awesome and wonderful and the Internet couldn’t work without them or other means of moving content around and stuff.
3. All that fuss about Netflix and Comcast (and Verizon and other carriers) is just about CDN stuff and Netflix not wanting to pay money it really ought to pay Comcast because, well, CDNs.
4. Don’t you feel silly now about that whole silly Net Neutrality thing, silly ignorant person?
There are a bunch of problems with this analysis. Notably:
1. As demonstrated by this
op ed raising similar arguments in 2010, content delivery networks (CDNs) and other ways in which the Internet does not “treat all bits equally” are not some fantastic new discovery that no one in the network neutrality movement has figured out. We all know about CDNs and other forms of prioritization imbedded in the network.
2. This does not violate network neutrality because network neutrality is not about “treating all bits the same” or other dumb ass strawman type arguments the anti-network neutrality folks would likethis to be about. This takes a basic high-level description that people use to illustrate the basic concept of network neutrality, and confuses it for the more sophisticated application of the principle. It’s like claiming that we can’t possibly have laws against race or sex discrimination in the workplace because ”equal pay for equal work” means “pay everyone exactly the same wage all the time for every single job everywhere” and then arguing how rigid application of “equal pay for equal work” ignores things like cost of living in your local area and would prevent merit pay raises.
3. The actual idea behind network neutrality is that last-mile ISPs, like Comcast, Verizon, AT&T or even scrappy little uncarriers like T-Mobile who are
trying to give you something free rather than leverage market power should not pick winners and losers by using their unique position as the access provider to the Internet to favor one application or service over another.
Or, stated more simply,
John Oliver is right. What we mean by “Network Neutrality” is “stop cable f#@!ery.” How things like the Comcast/Netflix fight and CDNs fit into this broader concept of “stop cable f#@!ery” or — for polite company — “network neutrality” has been covered fairly effectively by
Tim Lee over at Vox, and Stacey Higginbotham
at GigaOm. As to whether the Netflix/Comcast business fits into the existing network neutrality rules or needs to be handled as a potentially different sort of “cable f#@!ery” called “interconnection f#@!ery” (assuming we decide Comcast was wrong in the first place, a public policy question that remains unresolved at the moment pending more information), you can see conflicting views on that between
myself and my friend
Marvin Ammori.
Below, I demonstrate the fallacy of this “if you all just understood about CDNs and stuff you wouldn’t want this silly network neutrality’ thing” by comparing it with the two cases we’ve actually used the FCC to enforce network neutrality,
Comcast/BitTorrent and
AT&T/Facetime. I will then briefly touch on the ‘where does Comcast/Netflix’ fit in all this.
Testing The CDN Argument Against Reality.
We have, so far, two cases involving adjudication of net neutrality rules. Does this “it’s really about CDNs and anyone in favor of net neutrality doesn’t understand about CDNs” stack up against the two actual cases where a net neutrality process was invoked?
Comcast/BitTorrent
To refresh people’s memory, as Comcast explained after
the fact in a filing with the FCC, Comcast used forged reset packets to disrupt use of peer-to-peer applications for fileswapping, notably BitTorrent. The FCC in its
2008 Comcast Order found that this practice “blocked or degraded” use of a lawful application under the Internet Policy Statement, the precursor to the 2010 Net Neutrality rules.
Did Comcast/BitTorrent involve a CDN?
No.
Did the presence or absence of CDNs influence Comcast’s conduct in any way shape or form?
AT&T/Facetime
For those who do not recall this issue, back in 2012 AT&T Wireless announced it would not include Facetime on some iteration of the iPhone except at the highest service tier. Following the FCC’s rules then in force at the time, Free Press and Public Knowledge filed a formal notice that it believed AT&T was in violation of the network neutrality rules and used the formal complaint resolution process. AT&T and FP/PK agreed to a phase in plan. Complaint resolved.
Did this involve a CDN?
No. It didn’t.
What Do We Learn From This?
In the two cases we have to date that have actually used network neutrality rules, CDNs did not figure into it at all. The issue had to do with access by users to applications that they wanted to use. So, regardless of the whole CDN thing, if we don’t like broadband access providers limiting people’s access to the services or content or application, we need a network neutrality rule.
So where do CDNs and peering fit into this, I hear you ask? Well, for starters, if you use rules about CDNs and other forms of peering/interconnection as a fairly obvious proxy for targeting specific applications and services, that ought to violate network neutrality as well. And, of course, one can use these tools to achieve the same effect as simply blocking and degrading (or prioritizing) in the last mile.
Hey, Harold, Didn’t You Write Something 8 years ago that explained that CDNs and other methods of relieving congestion are a good thing and that absent Net Neutrality the incentive to use these techniques would diminish?
Right you are imaginary narrator construction!
In this blog post, I explain that CDNs and direct interconnection (along with other tricks) are part of “provider side provisioning.” Far from being a net neutrality violation, this “provider provisioning” is a very good thing. It creates a ”virtuous cycle” that stimulates investment in broadband infrastructure by edge providers and carriers alike.
But the incentive for content providers and others to build CDNs and similar infrastructure is diminished if broadband access providers can block, degrade or prioritize the last mile traffic. Why bother to ship stuff express if it’s going to sit three days extra on the loading dock unless I grease the palms of the broadband access provider that’s supposed to deliver it? Similarly, the incentive for broadband access providers themselves to invest declines when they can ‘monetize the scarcity’ by charging more for prioritization.
In other words, CDNs don’t violate network neutrality, CDNs are possible because of network neutrality.
Did I just blow your mind or what? I was gonna call this “what people who say ‘what net neutrality people don’t understand about the Internet’ don’t understand about Net Neutrality,” but I decided that was too long and kinda confusing.
Hey, That ‘Virtuous Cycle’ Stuff Sounds Familiar . . .
Yes, it was the rationale the FCC adopted in the
2010 Open Internet Order. Not that they gave me any credit for describing it in 2006, mind . . .
You mean that Order where they have a lengthy discussion about CDNs and all this other crap and explain why issues involving CDNs and interconnection are NOT covered in the Network Neutrality Rule they adopted because network neutrality is not about “treating all bits equally” or some other dumb ass reductio ad absurdum?
That would be the one.
And wasn’t that reasoning totally affirmed by the D.C. Circuit in Verizon v. FCC as supported by the evidence and rational and all?
And didn’t you write this long thing about how what is going in with Netflix and interconnection is an interconnection issue, not a net neutrality issue, and that the concerns here have a lot more to do with traditional interconnection issues than with manipulation of the termination monopoly?
Why yes I did. It’s this
thing over here. Although the case here is not entirely clear cut. It’s very clear you can use tools like interconnection and CDN policy to discriminate against particular types of content or applications and use them to violate the actual net neutrality rule, and the net neutrality rule should cover this kind of deliberate targeting. But even if it isn’t deliberate, interconnection issues raise a bunch of concerns, have a long history of being manipulated for anticompetitive effect, and have much broader implications for how the network as a whole operates.
So why do people broadly talk about Comcast/Netflix in the context of net neutrality?
Because what is going on with Netflix and net neutrality generally fall into the category of stuff John Oliver accurately described as “Stop Cable Fu@#ery.” For most people, “net neutrality” doesn’t really have a technical meaning. It means “stop cable fu#$@ery!” This does not mean they are ‘wrong’ or ‘stupid’ or don’t understand stuff. They understand very clearly: “We are being f#@!ed with. We don’t like it. Make it stop.” Telling them “technically, you are being f#@!ed by a CDN/peering dispute and not by paid prioritization in the last mile, silly person” does not really make them feel better about it or resolve the problem.
If that sounds odd, think about traditional telecom law. We have a general concept that the telephone system is supposed to ‘treat all calls and all customers the same.’ This isn’t literally true, as opponents of classifying broadband as a Title II telecom service
continually remind us. But it generally means the phone company can’t decide it will try to get people to order more pizza from Dominos or from Pizza Hut by deliberately dropping calls or degrading sound quality to rival pizza places. We also have a concept called “interconnection,” so that one telephone company can’t say to Pizza Hut “hey, you have to use my phone service and pay my ‘commercial rate’ to reach my phone subscribers instead of picking your own phone company that offers you cheaper rates.”
Both of these ideas are incorporated in the idea of Title II telecom service. But we break them out into separate rules (common carriage, interconnection) to implement the concept.
Which brings me back to Comcast/Netflix v.the 2010 “Open Internet” rule adopted by the FCC. From the perspective of the Comcast subscriber, it doesn’t matter what you call what happened to Netflix. For them (and everyone else worried about this), it is is all about what John Oliver calls “Stoping Cable F#@!ery.” When Joe or Jane Comcast Subscriber tell the FCC ‘protect the open internet and net neutrality’ they are not making a refined technical and legal argument on what rule should cover CDNs and interconnection.
Which is why all these technical arguments about how Netflix could or couldn’t control the traffic don’t really make much difference to the public. What people are really saying is:
“Goddaammit you miserable cable bastards! You just raised my cable bill two months ago. You make me rent a set-top box and keep upping the rental fee. You keep pushing me to rent a modem and then raise the fee on that. You just slapped me with a 300 Gig bandwidth cap and then you turned my wifi router, which you make me rent from you and charge me for, into an open hotspot.
“And now you are effing with my NETFLIX TOO?!!!”
So what is driving the 100’s of 1000’s of comments is not confusion about CDNs. “Net neutrality,” in addition to being a specific rule, is more broadly the way the FCC stops all the cable fu#$%ery. From the subscriber perspective, this is a clear example of cable fu#$@ery. Therefore it is a net neutrality violation.
So are you saying Netflix was right?
I’m also saying that, as a factual matter, there is remarkably little disagreement about what happened. i.e., Netflix switched in 2010 from using a CDN to using transit via Level 3. Comcast refused to upgrade the ports unless Level 3 paid CDN rates. Then Cogent offered to provide Netflix with transit. Netflix also offered a CDN service, but wanted to interconnect for free. Comcast and other broadband providers refused to upgrade the links under traditional transit terms, arguing this wasn’t suitable traffic for transit. Netflix traffic congested the transit links, causing steady degredation in the ability of subscribers to use Netflix’s service. Then Netflix cut a direct interconnection deal with Comcast. The quality of Netflix service on Comcast system jumped to equal that of services like Cablevision that directly interconnect with Netflix via Netflix’s CDN.
Whether Netflix was right or Comcast was right is a separate question that depends on how you categorize the behavior. If you categorize this as “Netflix deliberately chose to send its traffic to congested links to try to get free CDN-type transit,” then Comcast was right and Netflix are evil greedy bastards trying to offload costs and sleaze in traffic.
OTOH, if you categorize this as “Comcast is leveraging its market power to force Netflix to use particular routes if it wants to maintain viable service, denying high-volume customers to its transit rivals Level 3 and Cogent, and using its massive volume of customers to dictate the economics of the industry.” Then clearly Netflix is right and Comcast is wrong.
I will also note that we see similar arguments in telecom all the time around interconnection issues. Are T-Mobile and other competitors miserable parasites that refuse to build their own towers and backhaul by sponging off
mandatory data roaming and
special access? Or are T-Mobile and other competing wireless providers valuable competitors being shut out by incumbent providers leveraging their monopoly control of scarce spectrum and backhaul marketshare? Are CLECS (competing telephone companies) miserable parasites engaged in regulatory arbitrage by demanding interconnection rights for VOIP traffic? Or are they important competitors who need mandatory interconnection to make sure that competition can even happen in the first place?
Which is what leads me to my alternative way of looking at Comcast/Netflix that has nothing to do with right or wrong or CDNs or anything else people want to argue about. “Hey, this looks like an interconnection fight! We should really assess what the heck is going on here and figure out the implications, keeping in mind that the core universal values of our communications system that the FCC
voted on 5-0 back in January as universal values, aka what Chairman Wheeler calls the “
Network Compact,” tells us to evaluate this through the lens of protecting consumers and encouraging competition.
So far, however, I am pretty much the only person saying this last part other than Tom Wheeler.
Stay tuned . . .
Related Posts:
- S. Korea "Sender Pays" Is a Warning, Not a Model, or Why (Almost) Everyone Keeps Telling the EU This Is a VERY Bad Idea. by Harold October 14, 2022 Economist/NYT opinion writer Paul Krugman coined the term "Zombie idea" to describe an idea that, despite being repeatedly refuted with evidence, keeps coming back. Not…
- Why Canada's C-18 Isn't Working Out As Expected. by Harold July 24, 2023 Back at the end of June, Canada passed C-18, aka "The Online News Act," a law designed to make Google and Facebook negotiate with news…
- Get Ready for the 2022 Season of Spectrum Wars! by Harold March 15, 2022 It isn't the sultry Regency drama of Bridgerton, the action psycho-drama of Moon Knight, or even the, um, whatever the heck Human Resources is. But…
- My Insanely Long Field Guide to the Fox29 Philadelphia (WTFX-TV) License Renewal Challenge. by Harold August 29, 2023 In July, the Media and Democracy Project filed a Petition to Deny the license renewal of Fox29 (WTFX-TV) in Philadelphia. The Petition rests on a…
- Does SCOTUS EPA Case Impact Net Neutrality? Here’s Why I Say No. by Harold July 1, 2022 For most people, the Supreme Court’s decision in West Virginia v. Environmental Protection Agency was about environmental policy and what the Environmental Protection Agency can…
- AI Policy and the Uncanny Valley Freakout. by Harold June 30, 2023 We have been debating, on and off, about the issues around artificial intelligence and AI governance for some time now. Here at Public Knowledge, we…