FCC Drama on Net Neutrality and Spectrum
The February 15 FCC Open Meeting promises to be one of the more widely-viewed and discussed Commission meeting in a long time. The agenda includes some action on the upcoming (next year) Spectrum Incentive Auction (moving spectrum licenses from TV broadcasters to mobile networks) as well as action on the Open Internet notice, AKA the net neutrality rules. These issues have been widely discussed here and elsewhere, but there’s still a great deal of confusion about the underlying issues so here’s a note of clarification.
The Open Internet Notice of Proposed Rulemaking is the Commission’s response to the DC Circuit Court’s action on the FCC’s 2010 Open Internet Report and Order. The Commission’s order imposed anti-discrimination and transparency rules on broadband Internet Service Providers; the Court said the transparency rule was fine, but it found that the anti-discrimination rule was defective. In particular, the court said that once you classify ISPs as “Title I Information Services” you can’t very well impose the kinds of rules on them that are found in Title II of the Communications Act. The court agreed that with the Commission that it’s appropriate for it to read the Act as granting it vbroad authority to take actions to stimulate the construction of high-capacity broadband networks, but it didn’t like the way these particular rules were framed. The court said that the kinds of rules the Open Internet order imposed are simply inconsistent with Title I, so it could either modify the rules or change the classification of broadband services.
The supporters of net neutrality believe that the particular type of anti-discrimination imposed by the 2010 order is essential to the Internet’s progress and proper functioning, and the opponents believe that more sophisticated network practices could actually make the Internet work better. It’s not really unfair to say that net neutrality advocates see the Internet as no larger than the web, while critics see it as potentially working for a broader pool of applications, such as voice and video conferencing, massive use of video streaming by video services built on the Netflix model, and the Internet of Things, a chatty service in which devices talk to other devices without human participation.
It’s also fair to say that net neutrality advocates believe that diverse applications can work fine in the Internet of today, as long as homes and small offices can get higher capacity access and lower prices. Net neutrality advocates believe the Internet is a public utility, like water, gas, electricity, and waste water treatment, while opponents see it as an information system in which continual investment and upgrade is the norm. So this debate can be characterized as a dispute about how best to design a cheap, general purpose network that enables each user and every application to get the kind of service it wants and needs.
But it’s also a debate about how to spread the costs of building a network that requires continual upgrades or a network that is so incredibly fast and powerful that it rarely needs an upgrade. The latter line of thinking leads to the “Gigabit Cities” initiatives that are becoming increasingly common around the country, and the former leads to cost-sharing models where producers such as Netflix that account for disproportionate traffic loading relative to the user headcount help ISPs pay for the delivery of their content. Netflix now transmits more than 34% of the Internet’s traffic load during TV prime time in the US to less than two percent of its users (about 1.2%, actually).
The Netflix story becomes very interesting when the count of streaming video users rises by an order or magnitude. If 12 percent of the Internet user population is streaming at prime time, the traffic load goes up to 340% of today’s level; and if it rises to 60%, the load goes up to 1700%. If the resolution of TV images then goes up to the 4K level, the load would likely increase to 6800%. That’s a lot, but advances in compression probably reduce the load to something like 4 – 5000%.
Will gigabit networks – 2500 to 5000% faster than the typical residential Internet connection in America today – solve this problem? They certainly would, to the extent that the capacity bottleneck is located in neighborhoods. The trend in residential gigabit networks so far, however, is to grab bragging rights by provisioning homes with ultra-high speed connections but to provide only modest increases in capacity between the neighborhood network and the Internet switching centers themselves. Gigabit builders know that people are not going to use their connections at full throttle most of the time; in fact, gigabit networks don’t carry substantially more traffic that the more common 40 – 100 megabit networks provided by most ISPs. So the most you’re likely to see behind a gigabit connection would be an increase of two to ten times in the backhaul, if that. At the point where networks connect to other networks in the two dozen largest American cities, the standard form of interconnection is still a 10 Gbps Ethernet port.
So we’re not really dealing with the traffic increase we would see if cable TV switched to the Internet on a massive scale. Consequently, it’s probably wise to keep an open mind on business models for the time being.
On the spectrum side, take a look at my op-ed in today’s edition of The Hill. I point out that Congress has actually authorized the FCC to conduce spectrum incentive auctions, something it hasn’t done regarding net neutrality. Congress wants to address the spectrum shortage afflicting mobile networks and smartphone users. The National Broadband Plan tipped Congress off that this would happen, and it has. While mobile users are short of spectrum, the interest in Over the Air TV is waning, so it makes sense to complete the Negroponte Switch from a wireless TV network and a wired telephone network to to the emerging norms of TV by wire and telephone by air.
Before it can conduct the auction, the FCC has to decide what to do about two contentious issues, unlicensed spectrum and spectrum caps. The interest groups that lobby for unlicensed spectrum believe that a Wi-Fi like system can flourish in the relatively low 600 Mhz range that TV broadcasters occupy, but there’s very little reason to believe it can, as I explained in “How to Kill Wi-Fi.” But the Super Wi-Fi crowd has been more successful at lobbying the FCC for spectrum than they have at actually building networks, and I don’t see them stopping. I suspect their motive is to foreclose commercial use of spectrum wherever possible.
The spectrum caps issue is even more problematic, since unlicensed will most likely come out of the incentive auction from the leftovers. Sprint, T-Mobile, and some rural carriers want discounts for 600 MHz spectrum so they can round out their spectrum holdings despite having sat out previous low band auctions. These firms didn’t participate in the 700 Mhz auction that AT&T and Verizon used to increase their holdings, and now they want to be rewarded for refusing to invest in their networks. To add insult to injury, they’re both owned by deep-pockets foreign owners who could very easily afford to bid the market rate for spectrum. And to make things even sweeter, Sprint has twice as much total spectrum as either AT&T or Verizon, but they complain theirs is no good because it’s in the 2.3 GHz range. I suspect that the two most successful (AKA “dominant”) networks would be willing to swap some low band that 2.3 GHz because the high band is better in cities.
In any event, the atmosphere at the FCC seems to be extremely toxic these days, as some commissioners regard others – even members of their own party – as deceitful or dishonest. That’s what happens when net neutrality is on the table, and it’s extremely sad.