When discussing Net Neutrality, a phrase that often gets bandied about is “reasonable network management”. Just what, exactly, constitutes “reasonable” network management? While ISPs like Comcast and AT&T would have you believe that packet-spoofing and content filtering are reasonable practices, we here at Public Knowledge disagree. In accordance with the core principles of Net Neutrality, we believe that a reasonable network management technology is one that does not privilege, degrade or prioritize traffic based on the content, applications or services that it is associated with. But what would such a technology look like? This week, Comcast and Time Warner both begin testing network management techniques that don’t seem to violate the aforementioned guidelines. While these technologies are a step in the right direction, they also come bundled with a unique set of challenges.
Clearly, if ten percent of Internet users suck up 80 percent of the available bandwidth on broadband networks, as has been suggested, ISPs should be allowed employ some form of network management to ensure that a few bandwidth hogs don’t ruin the Internet for the rest of us. But how can a provider discourage users from overindulging in high-bandwidth activities without targeting specific applications? Time Warner thinks it has the answer: by doling out a monthly allotment of bandwidth to each user. In the company’s new pilot program, users who stay under their limit will pay the same set fee every month. Users who exceed their limit, however, will be charged for the additional bandwidth used—just like with a monthly allotment of mobile phone minutes.
Starting today, Time Warner subscribers in Beaumont, Texas will be able to choose from a number of different Internet service packages, ranging from a 768Kbps connection with a 5GB cap ($29.95 per month) up to a 15MBps connection with a 40GB cap ($54.90 per month). Any additional bandwidth used (either upstream or downstream) will be billed to the user at a rate of $1 per Gigabyte.
On the surface, these limitations seem somewhat reasonable. However, when you consider that Sprint lets users eat up to 5GB/month on its wireless data network, that 5GB doesn’t seem quite as generous for a wireline connection. And even the “high end” 40GB data cap could prove inadequate. As Ars Technica duly notes, “The Internet is increasingly being used as a vector for distributing software and digital video content and also facilitates multiplayer gaming, video conferencing, real-time collaboration, interactive remote desktop access, file backups, and many other bandwidth intensive activities.” Despite what you think, that 5GB might disappear fast—especially if you have video game consoles, video streaming appliances and online backup services hooked up to your network—as an increasing number of us do.
So, might there be another way to regulate traffic by punishing only the heaviest users while leaving the rest of us to stream video in peace? Comcast seems to think so. In the wake of the controversy caused by the company’s Bit Torrent blockade, Comcast has begun testing a new method of network management that identifies which users are eating up the most bandwidth and temporarily imposes limits on those users during times of network strain. According to a FAQ published by Comcast, “the technique measures only aggregate bandwidth consumption, not the protocol or content being used by customers.” During times of network congestion, the technology “manages [the bandwidth usage of users] until their usage falls below established bandwidth usage thresholds or until network congestion ends.”
In theory, this method sounds a whole lot better than the techniques that are currently being employed by Comcast. However, given the fact that the company previously promised that it only managed traffic during times of congestion, only to have those claims disproved, we might have to take the term “network congestion” with a grain of salt. Also unsettling is the fact that the company hasn’t revealed any further details about what constitutes a high-bandwidth user and how limited those users will be during times of congestion. If Comcast chooses to roll out this type of technology, it will have to be upfront with its customers about what the exact limitations are. And given the company’s secrecy throughout the Bit Torrent fiasco, I wouldn’t hold my breath for that kind of transparency.
So, we’ve established that while technically “neutral,” both Time Warner and Comcast’s new network management techniques are not without their share of issues. There is still, however, one very large elephant left in the room: the fact that both Comcast and Time Warner are cable television providers. And as we all know, despite the industry’s constant invocation of the P2P bogeyman, at present, the largest bandwidth hog is actually streaming video. Clearly, the emergence of online video is something that cable video providers find very threatening and by capping off bandwidth usage, they’re effectively killing two birds with one stone; discouraging users from using their Internet connections for video while increasing the efficiency of the network. Is this anti-competitive? It sure seems like it. But is it anti-neutral? Technically, no. While Time Warner and Comcast both deliver video and Internet service via the same pipe, the two services live on separate networks.
Even that could change, however, depending on how well the cable industry responds to the coming threat of IPTV. While unlikely, it’s possible that both Time Warner and Comcast will branch out into video over IP in the coming years, if such a move is dictated by market pressures. And then what? Will Time Warner count videos streamed using its service when calculating your monthly bandwidth usage? Will Comcast penalize you for purchasing its video content online? If the answer to these questions is ‘no,’ then the two companies will be prioritizing their own video content over video content from other providers on the same network. In which case, we’ll find ourselves back at square one.
SOME
SOME CLARIFICATIONS
Bandwidth capacity is a measure of momentary (instant) flow use of Gigabits and should not be confused with Gegabyte use over time. Reducing congestion in peak periods is about reducing coincident Gigabit use which may come from large or small Gegabyte users who contribute to congestion on the same per Megabit basis, regardless of their total Gegabyte use over time.
If Comcast restricts only “heavy Gegabyte users” in peak congested periods, then small peak users are getting subsidized by “heavy users” on a per Megabit basis because both are causing the same amount of congestion on that basis.
If Comcast insists on centralized non-price rationing methods to reduce peak congestion for the artificial shortages it has created, at least bandwidth should be reduced for all peak users on some proportionate basis - most likely in proportion to their maximum connection rating. Otherwise, it’s not neutral management of congestion in terms of reduced use based on causality - in conflict with its newly claimed “agnostic protocol management”.
Otherwise, the term “bandwidth hog” is a misnomer, arising both from propaganda to undermine net neutrality as well as the practice of overselling bandwidth in terms of the Gegabytes produced by a connection, then declaring large users as over the line - while simultaneously refusing to draw lines - until now with metered Gegabytes, which is a crude method of rationing bandwidth since off-peak use does not affect congestion.
By continuing to cherry pick away at “heavy users”, the network providers put their market power and strategic pricing control front and center with selective congestion controls and absurd low limits on Gegabyte use with sharp overage penalities. Any competitive market would have a full range of available Gegabytes and bandwidth tiers in Megabits along with multiple levels of service quality and peak/off-peak pricing plans.
Meanwhile, the term “bandwidth hog” should to be challenged everywhere it appears - it represents the essential justification for establishing fast lanes for selective, high-price content versus everything else that gets dumped into the slower bus lanes, the whole grand design touted to “alleviate congestion from the coming exaflood”. What it really means is higher prices will be assessed to escape the artificial shortage, existing service will be degraded and much produced content will be denied access to the fast lane.
There are no “bandwidth hogs” - there are users of different amounts of available maximum Megabits in designated bandwidth tiers, and when they use (acquire) more bandwidth, they pay more and vice versa as explicitly provisioned in the business model of network providers.
Because network providers were caught flatfooted by the increased use of Megabits per connection that they explicity marketed and sold, is not a reason to blame the outcome on “hog” customers, particularly given the intentional underinvestment - it was easy enough to downgrade the connection capacity instead, to be consistent with an uncongested network, but they refused and kept overselling it.
Now that they’ve created the perfect storm of a congestion problem that must be solved by means other than investment or appropriate pricing, the only thing standing in the way of course, is … net neutrality.