“We can’t have a situation in which the corporate duopoly dictates the future of the Internet and that’s why I’m supporting what’s called net neutrality.” — Barack Obama, podcast, June 2006
[June 19: So much for pruning – added 300 words in corrections and background.]
On Friday, June 13, FCC Chairman Tom Wheeler made a short but dramatic statement headlined Broadband Consumers and Internet Congestion. Though barely 450 words long and presented outside any formal setting, Wheeler’s reaction to the public hue and cry over the reliability of retail broadband in the US marks an important step forward for end-user welfare. His announcement puts the lie to the vehement criticisms levelled at him about his betrayal of the Open Internet concept (the FCC’s term of art for net neutrality).
Many of his critics also assumed that the Wheeler FCC would never look into paid peering arrangements – well, they actually said they wouldn’t (“… the rules we propose today reflect the scope of the 2010 Open Internet Order, which applied to broadband provider conduct within its own network.” NPRM, fn 113 – pdf uploaded here). That is what Wheeler has now directed Commission staff to do (request “information from ISPs and content providers”).
While the American public are clearly confused by the net neutrality debate, and for good reason, many ISP subscribers have begun to question whether they’re getting the broadband they’re paying for – whatever the underlying business and technical issues may be. Excerpts from Wheeler’s statement follow (the full pdf is uploaded here):
“For some time now we have been talking about protecting Internet consumers. At the heart of this is whether Internet Service Providers (ISPs) that provide connectivity in the final mile to the home can advantage or disadvantage content providers, and therefore advantage or disadvantage consumers. …
“Consumers must get what they pay for. As the consumer’s representative we need to know what is going on. I have therefore directed the Commission staff to obtain the information we need to understand precisely what is happening in order to understand whether consumers are being harmed. …
“The bottom line is that consumers need to understand what is occurring when the Internet service they’ve paid for does not adequately deliver the content they desire, especially content they’ve also paid for. In this instance, it is about what happens where the ISP connects to the Internet. It’s important that we know – and that consumers know.”
Here’s how convoluted the coverage of net neutrality has become.
On May 22, Margaret Sullivan, the Public Editor of the NY Times, responded to vehement reader complaints about the paper’s coverage of the now-infamous FCC vote on new Open Internet rules. In a piece entitled “Readers Say a ‘Net Neutrality’ Vote Was Reported Upside Down and Backward,” Sullivan noted that the previous week’s coverage of the vote brought
“… many protests from readers, and from outside commenters, who say the new development was presented in a misleading way, one far too accepting of the perspective of the F.C.C. and its chairman, Tom Wheeler, who has spoken in support of the proposal.”
The editor’s note reflects the remarkable polarization in reactions to Chairman Wheeler’s public comments about neutrality – all the more so given how polarized feelngs already were. The Times staffers themselves seemed as perplexed as everyone else about what’s good or not for the Internet. To the point that Sullivan used her bully pulpit to side with readers who thought Reuters and the HuffPost had a more accurate take (“F.C.C. Votes for Plan to Kill Net Neutrality”) than the Times’ own Edward Wyatt (“F.C.C. Backs Opening Net Neutrality Rules for Debate”).
“A clear strike against the commonly understood idea of net neutrality”
In trying to make sense of these views, I would start by questioning the idea that there is anything resembling a “commonly understood idea of net neutrality,” in Sullivan’s words. The whole setup to John Oliver’s hilarious parody on his Last Week Tonight show – against “cable company fuckery” – is based on how quickly regular folks glaze over at the mere mention of network neutrality, let alone details only a policy wonk could love (see his video rant at the end of this post).
Still, Oliver not only manages to break through the boredom, but he is also credited with prompting over 45,000 irate consumers to leave comments on the FCC site – so many that it crashed (not to mention the 3.7 million views Oliver has had on YouTube). Chairman Wheeler responded to this unprecedented surge of public interest by announcing on June 13 that the FCC is going to review interconnection agreements after all. Not bad for a guy whose appointment as FCC chair Oliver compared to choosing a dingo to do your babysitting. (To be fair, Oliver’s people should also have included incumbent telco fuckery in the script.)
The sudden surge of interest in net neutrality prompted many in the mainstream press to offer basic how-it-works explanations. Some went so far in trying to dumb down the technology – without the Oliver hilarity – that they produced distinctly unhelpful analogies, like the one in the screen grab below from the NY Times.
Explaining the role of ISPs in packet transmission by referring to them as “sorting facilties” makes it sound as though Comcast (or Rogers) is just another Netflix hub – like a Fedex parcel facility – instead of the vertically integrated gatekeeper that subscribers pay every month in the often vein hope of seeing their favorite videos run in HD without constant buffering. That control not only confers tremendous market power; it also confers a huge conflict of interest, to the extent that the incumbent ISPs have content to sell that competes directly with edge providers – especially OVDs like Netflix.
Even Wikipedia is getting hammered by so many attempts to flesh out the story that the net neutrality entry is currently running the disclaimer that it “contains too many or too-lengthy quotations for an encyclopedic entry.” Rich with detail, the entry offers sub-sections on some of the concepts that have emerged in the public debate (open Internet), along with others too wonkish for public consumption (common carrier, dumb pipe, end to end principle). I also notice that it was updated as recently as yesterday.
A basic broadband consumer issue, tangled in technology
It’s unfortunate the American debate got kicked off by such a volatile and readily misunderstood chain of Internet-related developments. Hard on the heels of the FCC’s loss to Verizon in the DC Circuit Court of Appeals over its Open Internet Order, Netflix and Comcast signed a deal in February to eliminate the interconnect middle man. Netflix now sends its traffic
right to the Comcast local access network to interconnect points where (I assume) it can simply hop across a switch to reach Comcast’s local access network – a move that has improved end-user performance by an appreciable amount. But not what you’d call dramatic: from a mean of 1.68 Mbps in February to 2.72 in May, with a rank increase in this period from 11th to 5th (see Netflix blog).
This deal kicked off a major shouting match over a concept I’m pretty sure I had never encountered before: paid prioritization. One of the most widely followed advocacy groups, Free Press, summed up its reaction to this idea on its Save the Internet site by calling paid prioritization “The Antithesis of Openness on the Internet.” Imagine the further surprise, therefore, when, barely one month after signing the deal, Netflix CEO Reed Hastings took to the company blog in a tone that gave the advocates a run for their money.
Under the headline Internet Tolls And The Case For Strong Net Neutrality, Hastings blasted Comcast and the other big ISPs that were taking advantage of their market power to screw everyone else – Netflix included:
“Without strong net neutrality, big ISPs can demand potentially escalating fees for the interconnection required to deliver high quality service. The big ISPs can make these demands — driving up costs and prices for everyone else — because of their market position. For any given U.S. household, there is often only one or two choices for getting high-speed Internet access and that’s unlikely to change. Furthermore, Internet access is often bundled with other services making it challenging to switch ISPs. It is this lack of consumer choice that leads to the need for strong net neutrality.”
Adding to the mystery of this recantation, the Netflix-Comcast deal is buried deep under an NDA. Nobody but the principals knows how much Netflix paid, why it came to the table when it did, exactly how long the agreement lasts, or how much the OVD is saving over whatever it was paying its CDN or other middlemen. The closest I’ve seen to a plausible cost estimate appeared in The Verge in a piece on Hastings’ “Deal with the devil“…
“Wedbush Securities estimates Netflix is paying between $25 million and $50 million a year to Comcast, and industry analysts like Frost and Sullivan’s Dan Rayburn think the cost to Netflix could be as low as $12 million a year. That estimate is supported by the fact that Netflix has not filed any notice with the SEC altering its guidance for this quarter’s financial performance…”
The writer concludes on this evidence it was likely “Netflix was finally able to extract terms from Comcast it found favorable enough to agree to.”
The highly alarmist talk about risks to net neutrality took an even worse turn starting April 23, when the press got hold of a “leaked” version of proposed changes to the FCC’s rules for the Open Internet. The ideas being mooted became part of a notice of proposed rulemaking (NPRM) after the May 15 vote among the five commissioners – the standard procedure for soliciting public comments. Equally true to form, the vote split along party lines, with the two Republicans delivering themselves of fulminating dissents about the horrid consequences for any further attempts to “regulate the Internet.” Commissioner Pai, for example, ends his dissent – a closely argued, 5-page document – by saying the following (pdf here):
“… undertaking such a ‘politically corrosive’ rulemaking on dubious legal and policy grounds will swamp what should be an independent, expert agency with years of litigation and partisan division. That is not good for broadband deployment, that is not good for consumers, and that is not good for future of the Internet.”
Wheeler has done something very smart as of last Friday, at least in the view of most of Washington’s open Internet advocates. He has reframed a very technical discussion in terms that have thrown the spotlight back where it belongs: on the millions of ISP subscribers who have not been getting the broadband service they’re paying for. But there’s still plenty of reason to be pessimistic about American (and Canadian) ISPs getting persuaded or obliged to provide what they say they’re providing – unfettered access to legal content on the public Internet.
Here are four issues I think may stand in the way of achieving that goal:
1. Whatever you call “neutrality,” it’s still way too technical. We’ve seen the debate get run off the rails over the concept of paid prioritization. Regulators and the various parties – especially the ISPs now under assault – are going to be talking about peering, transit and paid peering, about interconnection in general and content delivery networks in particular (coming soon to an ISP near you with the Apple logo plastered all over it). Even worthy and non-technical ideas like innovation without permission are not familiar or comfortable to mainstreamers. The more bafflegab, the worse for consumer welfare, where ignorance is never bliss.
2. The whole debate remains remarkably emotional and polarized. The incumbents and the politicians they pay off in Congress are not going to back down over control of the gates. Opponents of regulation to achieve greater equity in the retail broadband market won’t be drawn into a real discussion. They talk as though they want the same goals – like innovation and free speech – while being adamantly opposed to what they claim are e.g. “19th century ideas” like making ISPs common carriers (the Title II gambit). It’s hard to debate with powerful idiots.
3. Whatever the FCC does, not much may change in the short term. Just like the CRTC, the FCC not only has severe constraints on its jurisdictional freedom of action. It also has to contend with the prospect of getting dragged back into the courts all over again, possibly for years – as Comcast and Verizon have already done, to good effect from their perspective. In one of the most balanced accounts I’ve read of these legal tangles, Kevin Werbach and Phil Weiser wrote a few weeks ago in the Huffington Post that a great deal depends on the next step after rulemaking – to wit, enforcement: “The rules get all the attention, but the real question is what the FCC does with them.” (The NPRM does contain a brief comment on enforcement at para 173.)
4. Buying broadband isn’t like buying gas: mileage always varies. Over the last three or four years, policymakers in other jurisdictions, like the UK, have paid careful attention to the inevitable advertising slogan in this market: you’ll get “up to” a certain amount of bandwidth, depending on a whole host of factors. The truth is that some of these factors are out of the ISP’s control, at least in day-to-day network management. Packets have overhead and overhead isn’t payload. Big spikes in traffic will happen around unusual events, like the World Cup. Many subscribers have bottlenecks in their modems or computers they could fix to get more bandwidth. And so on.
That said, there is absolutely no doubt the incumbents in the US and Canada have used the “up to” language in an unaccountable fashion, to which subscribers have no recourse (one of the best reviews of this problem is the 2013 PIAC study I’ve mentioned before: Transparency in Broadband Advertising to Canadian Consumers – pdf here). To its credit, Ofcom has been all over this problem for years. Back in 2010, it posted speed measurements taken in conjunction with Sam Knows and had this to say:
“In April 2009, average actual (or download) speeds were 4.1Mbit/s, 58 per cent of average advertised ‘up to’ speeds (7.1Mbit/s). In May 2010, average download speeds were 5.2Mbit/s, 45 per cent of average advertised ‘up to’ speeds (11.5Mbit/s).”
There’s one more critical issue behind the “mileage varies” problem: insufficient investment in network upgrades. The incumbents are encouraged by Wall St. and their investors to avoid spending money wherever possible. If by doing so they are allowing serious congestion to interfere with end-user enjoyment, then Wheeler may have found his rallying cry – though not if the Republicans have anything to say about it.
Congestion gets real
What an irony that would be. Network congestion was a made-up issue that Bell Canada sold to the CRTC four years ago so it could prejudice the operations of its wholesale ISP customers, through both throttling and usage-based billing. But the growth of traffic in the meantime has been relentless. In its regular Zettabyte Era report, updated on June 10, Cisco says North American IP traffic will grow at a CAGR of 20% over the 2013-2018 forecast period. When we look at the VNI interactive page and filter for Canada, we find Internet video traffic alone will grow 3-fold from 2013 to 2018, a compound annual growth rate of 28%.
What’s happening in Canada is tracking growth in the US very closely: US video CAGR is 27% to Canada’s 28%. It also appears the huge chunk of peak-time downlink bandwidth taken up by Netflix streams in the US (over 30% of the Internet total) runs about the same in Canada. The big question is whether the congestion that has Chairman Wheeler worried is typical of how the incumbents across North America are running their operations. And of course whether there’s anything that can be done to get the incumbents to keep upgrading their bandwidth capacity without sending the bill to either the edge providers or their own customers.
In a subsequent post, I’m going to try to figure out what the American broadband crisis means for Canada and the CRTC…