Net Neutrality - (why) do we care?

In this post I hope to explain the newly famous phrase "net neutrality" to non-technical people who are interested, like politicians and journalists.

It is a challenge to properly define net neutrality. Everyone - Internet providers, tier 1 carriers, websites like facebook, twitter, services like content delivery networks (CDNs) - are talking about it, everyone of them mean something different. Of course I also have my own view of what net neutrality should be, but I will first try to explain the different points of view while remaining as unbiased as possible.

The Two Options

When you bought access to the Internet in 2007, you probably got a package including Internet access, an email address at your Internet provider and maybe some webspace where you could host your personal web site. The physical access was provided by a technician who would either use surplus wires from your telephone connection, unused frequency bands from your telephone connection or a completely separate set of wires.

When you buy access to the Internet nowadays, you will not only get Internet access and an email address, you will also get access to video on demand, IP-based telephony, IP-based broadcasting, cloud storage and more, either automatically included in the contract, optionally, or from different service providers or content providers. All of these services use the same infrastructure: the Internet. Depending on your Internet connection, you might not be able to watch TV while your roommate uses a video on demand (VOD) service. There are two main points where congestion can occur: your downlink and an Internet interconnection point (IX). If your Internet bandwidth is really poor, like 5 Mbit/s, the congestion is probably on your downlink, otherwise it is in an IX.

These two cases are completely different, even though the outcome is the same for the customer. Even today, where you mostly do not just buy Internet access, but also access to several other services, contracts include the speed at which the Internet connection should work. However, since improving the bandwidth on the "last mile" is very expensive - it amounts to laying fiber or cable to each and every home - Internet providers avoid it and instead sell Internet access with a speed of "up to x Mbit/s". A clause like this is often problematic because it does not state where the reason for the slow-down is when the maximum speed is not achieved. It could be that the signal quality at your home is poor, but it could also be that your Internet provider is not connected to the rest of the Internet well enough.

The point of view of the consumer is that they paid for access to the Internet, or, to be more precise, transmission of content across the Internet. In their contract, their access to the Internet might be limited by a maximum speed which might depend on how well the data signal can be received at their house (this mostly means how close they are to the closest distribution point), or a maximum speed that is enforced by software at the Internet provider, and in rare cases also by a total volume of traffic that can not be exceeded (without paying extra). There should not be anything else limiting the access to the Internet.

Note that the consumer did not pay for access to "Internet Provider X's Network", they paid for access to the whole Internet. However, the Internet is a best-effort service, as are the protocols at its core (TCP/IP). Effectively this means that some sites/services on the Internet might not work as well as others. This can be due to the distance between server and client and network congestion. When the problem is network congestion, especially when it is at the interface between the Internet provider and tier 2 or tier 1 carrier networks, there is the problem of who has to pay for upgrades of the interconnections.

The problem of net neutrality from the consumer's point of view is therefore: What are the responsibilities of the Internet provider? There are two competing points of view here:

  • Option 1: The Internet provider is paid for providing Internet to its customers. If the provider notices that a certain interconnect is congested and its customers can therefore not use the service as promised in their contracts, the Internet providers need to take steps to upgrade the congested interconnection.
  • Option 2: The consumers are using a certain service or a certain number of services which the Internet provider is connected to via a tier 1 or tier 2 carrier. These services use up a lot of bandwidth and therefore the content providers or the top tier carrier should pay for upgrading the interconnections and the increased cost of maintaining them.

The difference in these points of view is exactly, who pays for the faster connection? In the first case, the consumers pay for the better interconnection via higher fees for their Internet access, because the Internet provider needs to invest in better interconnections. In the second case, the carrier networks or the content providers themselves need to pay (which is actually the same, with one "hop" of difference). Of course, in the end, the consumer pays for the interconnection again, by paying more for the content which the content provider makes accessible.

This is why, in the second case, AT&T can promise to lower your bill, but they do not tell you that you would pay more for the services which need to buy better access to your Internet provider's networks to be usable.

"But," I hear you say, "if, in the end, it is always the consumer that pays, it does not matter which way it is done, right?"

As always, it is not quite as simple as that.

Internet Providers vs. Content Providers

Remember how I told you that some of the services included with Internet access or offered additionally by your Internet provider are provided by that Internet provider themselves? This creates a conflict of interest: Why, for example, should an Internet provider, who also provides a video on demand service, improve the interconnection to a tier 1 carrier network that is severely congested because a rival video on demand service is heavily used by the Internet provider's customers?

The rival's video on demand service might "feel" better, be easier to use, have a larger library of videos or be cheaper for consumers than the Internet provider's own video on demand service, however, since the interconnection is severely congested, it is almost unusable. The Internet provider can then tell customers to use its own video on demand service and demand payments from the rival video on demand service for better interconnection. Something like this has already happened, even though Comcast denies that it refused to pay for upgrading the interconnections to give their own service and advantage.

And in the end, the consumer always pays - but not necessarily the same. With option #1, the Internet provider needs to pay for the better interconnection themselves and ask more from their customers, while with option #2 the service fees for Internet access stay the same, but Internet providers can ask whatever they want for their own video on demand services and make sure that they stay competitive by asking enough for faster interconnection from other content providers to make sure that these other content providers are no threat to their own business.

So in the end, I believe the consumer might pay less in the first case. Note that with option #2, content from content providers like video on demand services gets more expensive for everyone in the long term, not just customers of the Internet provider that charges the content provider for faster interconnection. Overall, this favours content providers run by the same company that happens to provide you with Internet access. You might have to change your video on demand subscription, your TV-broadcast-over-IP provider, your voice-over-IP provider etc. every time you change your Internet provider (which might happen because you move to a new home), just to get acceptable service: no dropped calls, no stuttering video.

The situation is very similar if there is no conflict of interest, i.e. when your Internet provider does not run its own video on demand service. When option #2 is chosen and the Internet provider gets paid by the content provider to upgrade interconnections, there are still two problems:

  • Access to the network is granted to those content providers who are able to pay the highest fee. A startup with limited budget might not be able to get a connection of sufficient quality without buying faster interconnection, effectively removing them from the competition. The Internet provider therefore chooses which services you can actually use on the Internet (in the sense of, which are usable: download speed, jitter, stuttering video...).
  • If the content providers need to pay the customer's Internet providers for faster interconnection, what are consumers paying the Internet provider for? Where does the money actually go?

Remember that Internet providers promise their customers Internet access up to a certain speed. If the Internet provider is not connected to the rest of the Internet well enough to be able to actually provide this speed almost simultaneously to all their customers (as it is the case in peak hours), then - this is my opinion - this is a breach of contract (promising something you can not actually do) and customers could sue the Internet provider, even without any net neutrality rules.

Why does this not work? Remember when I told you that Internet providers sell access to the Internet using a formulation like "speed up to x Mbit/s," and that this is a problematic formulation? This is where it kicks in - if a customer were to sue their Internet provider, they could just point at that clause in the contract and explain that the Internet is an "effort based service" and that, sadly, there was no effort from content provider X to improve the connection.

Transmission vs. Content

In my opinion, a lot can be learned from the way the European energy (gas and electricity) market is regulated. There is a strict distinction between transmission networks and power generating companies (the "content providers"). In Europe, you can change your gas or electricity provider easily, sometimes even on the Internet. A few clicks and the only thing that happens is that next month, the bill in your mailbox has a different logo and different numbers on it. This is only possible because of the Third Energy Package, which requires separation of energy transmission companies from energy generating companies.

A requirement like this completely prevents scenarios like the one above, where a transmission company can effectively decide which content providers consumers have access to and thereby which content providers can stay in the market. Aside from the conflict of interest situations, it is also the transmission providers which must ensure that there is enough capacity. For example, if a lot of people want electricity from a certain supplier, it is the duty of the transmission network to ensure that it can meet the demand. It is not the power plants that need to build additional connections to the individual transmission networks - power plants, like content providers, are always connected to the tier 1 network with sufficient "bandwidth" (imagine if a 500MW power plant was only connected with a 50MW power line, or if facebook was only connected to some tiny Internet provider from the arctic that is only connected to the rest of the Internet with 2Gbit/s).

In my opinion, this is what net neutrality is really about. "Net neutrality for the last mile" therefore just does not make any sense, just like a net neutrality for the last mile for electrons or methane does not make any sense. It also does not mean that a company can not sell Internet and services that depend on the Internet - there just needs to be a regulator that makes sure that it is actually consumers who chose their content provider, and not the Internet provider who chooses for them.

© 2010-2015 Stefan Birgmeier
birki@21er.org