Network neutrality is about control

The following is the text of a letter that I wrote in response to an article in the Washington Post this morning. The article asserted that pro-net neutrality groups' call to "keep the Internet where it belongs -- in the hands of its consumers" is "empty fluff that is obviously designed to incite rather to inform." I disagree, and here's my primer on net neutrality to tell you why.

Whereas the rhetoric is certainly intended to incite and not to inform (I don't believe that either side of any argument uses a slogan to inform, by the way), it isn't just empty fluff. In particular, as you noted earlier in the piece, the costs for services will eventually be borne by the customers regardless of who they pay for them and how they're provided. The purpose of the net neutrality folks is to put the control of paying for service in the hands of the consumers, where a customer who wants to use any number of services that consume large amounts of bandwidth pays for a higher quality, higher bandwidth, and probably higher-priced connection. This does two things: first, it makes the payments transparent, since the users know they are using bandwidth to watch television or movies, or other things that either take a long time or involve a high level of interactivity; second, it provides an avenue for innovative, high bandwidth services to get a start in the world.

If net neutrality loses, Google and Amazon will have to decide if it is more important for them to have high-speed (or more likely low-latency) access to consumers on each network that charges them separately for bandwidth utilization. More importantly, though, innovative new services (think things like YouTube and almost any AJAX-based service that is yet to be invented) and low-volume providers (think independent film makers, bloggers who post video of current events, etc) will be stuck in a position where their content may be discriminated against in a manner that they can't appropriately remedy. In particular, if they are too small, a company like Verizon might not even return their calls even if they wanted to get higher quality connectivity.

As it stands, it appears that players like Verizon are more likely to shut down services that they are more likely to get money from than from services that they haven't yet heard of, but without legislation, there are no guarantees that will be the case.

Who pays for the Internet

As you correctly point out, companies such as Google, Apple, Amazon, and Microsoft are already paying a large amount of money to have their content moved around the internet. In fact, a misconception about the way the internet works is that companies like these pay to have content put on the internet in some location near their facility and that the phone and cable companies then pay to drag it all the way across the country and into the end-user's home. This could hardly be further from the truth. In fact, the very large companies (such as those above, and thus the ones that are using the vast majority of the resources the telcos are looking at levying fees on) position servers near to where the end users live in order to provide the highest quality of service. This use of what is commonly referred to in the industry as CDNs (Content Delivery Networks) is costly and delivers the content to a point very close to where the ISP connects to their upstream provider (most, if not all local ISPs do not run nation-wide or world-wide networks, they buy their access to the internet from one or more backbone providers). This expensive delivery mechanism provides high performance by positioning the data as close as possible to the point to which the content is intended to be delivered. Companies such as Akamai have made a name for themselves providing this kind of service. On the really large side of things (Google as an example), some of the providers have their own CDNs that they have built by deploying multiple data centers geographically.

Generally speaking, a business customer buys internet access from either a local provider (such as Verizon) if they are a small customer, or a national provider (such as UUNet, Sprint, Global Crossing, etc) if they are moving a large amount of data. Most significant services purchase multiple, independent network connections to be able to provide the best service to their customers, and a select number of these either deploy multiple data centers or use CDNs to put content even closer to the "edge".

In the current scheme of the internet, content providers pay to have a certain amount of bandwidth available for transmission (and reception, but that's usually insignificant for content providers) and content consumers pay to have a certain amount of bandwidth available for reception. In the middle, the network providers (ISPs such as Verizon, Comcast, etc as well as UUNet, Sprint, etc) pay mostly for their own links to each other and within their networks. Due to the sheer volume of traffic in and out of their networks, this can amount to significant bandwidth, but is usually very inexpensive and in some cases can be free due to agreements with the upstream providers.

Further, some CDN providers, such as Akamai, provide content using servers installed on ISPs networks. In these cases, the content is cached on the Akamai servers and the ISPs only pay to receive each piece of content once. The equipment is housed at the ISPs location, but is paid for and managed by Akamai. As such, the bandwidth consumed by these services is local to the ISP (only on the ISPs internal network to its customers).

Internet Utilization

As a former owner of a consumer ISP and a long-time Internet user and network administrator (I started using the 'net in 1985 at NCSA, where I co-authored NCSA Telnet, an early communications program for what is now called the Internet), I also have seen the internet from multiple points of view. From my involvement in ISPs I understand how tight the margins are and how competitive the field can be, especially when dealing with consumers. I also understand that in most consumer networks (including ones that prohibit the practice), peer-to-peer communication (bittorrent, kazaa, gnutella, etc) consumes much more bandwidth than the content served up by Apple, Google, Yahoo, Microsoft, and others who are usually painted as "big users". This portion of the network doesn't have any well defined commerce going on and thus is somewhere that ISPs have to provide access to based on the fees they directly charge users.

Charging for commerce, not access

On today's Internet, sending and receiving data has already been paid for and what the ISPs that are resisting net neutrality are calling for is the ability to charge content providers a second time for access to their customers. An apt analogy would be the phone company attempting to take a percentage of any transaction that was done over the phone. The calling party has already paid for the phone call, the receiving party has either paid for the phone call (metered services or cell phone) or has paid for unlimited inbound calling through a subscription. However, the phone company sees that there is money being made by others transacting business over their phone lines and decides they deserve a cut.

Why is this a regulatory issue?

Most debates about how much to charge for things can be resolved at the commercial level. Companies like Google really could tell Verizon/Comcast/whomever that they'd just turn off access to customers on their networks instead of accepting intentionally reduce performance. However, this will only work for acknowledged leaders in the market, such as Google and Yahoo. Even companies like Apple are unlikely to be able to withstand taking such a stance alone against the phone company, and banding together would likely result in an antitrust suit by the phone companies (an ironic twist of fate that).

As with all basic utilities where the government has granted sanction or monopoly powers (and especially with telecommunications where taxes and fees paid for the construction of at least part of the infrastructure), Internet Service Providers need to be dealt with as companies whose customers do not have the power to exercise choice. For example, in order to get high-speed internet at my home, I can choose between the cable company and.... well, I can only choose the cable company, since the phone company doesn't provide even basic DSL to my location. There is no option to switch to another provider, no provision for dissent, and no avenue for redress other than the PUC and the federal regulatory agencies or congress. For small content providers, who are attempting to provide content and services to a nationwide audience, the situation is even more bleak. Nothing less than federal regulation is likely to provide them with a negotiating position.

And a word about censorship

Right now, ISPs are protected by special status granted to them in the DMCA and CDA (section 230), and therefore are not responsible for the content that is transported, only the transportation of that content. If net neutrality were to lose, the carriers would be free to decide (on whatever basis they choose) which content to charge more for. This means that they can charge more for big content (such as movie downloads), low-latency content (such as VIOP telephone calls and online gaming), or content that they believe is lucrative (pornography) or distasteful (political views they don't share). Without some form of regulation over this behavior, the protections for these ISPs must be removed as they will be exercising some form of editorial control over content.