Commentary: Some Data Packets are more Equal Than Others
The most important voice on the Internet is the American consumer - or so you'd think if you visited www.savetheinternet.com. The efforts of the creators of www.savetheinternet.com are laudable, but the mood of the contributions to the debate leave me with the impression that the contributors are just "looking for a fight" rather than actually making any sensible suggestions.
One of the problems is that the site is hosting two separate debates under the same theme of "‘net neutrality."
Debate one concerns whether it is even possible for network operators to "influence" Internet users' access to various websites, and therefore "influence" uptake of ideas and services provided by such sites. "Influence" is universally argued as that of being able to "slow down" access to one website or service provider in favor of another who may instead be commercially or politically engaged with the network operator.
Debate two regards whether all services and information on the internet should be treated equally, and whether therefore network operators should give every packet of data that they transfer equal priority.
While at first these two arguments seem to amount to the same issue, this is not the case.
Debate One - Is Content King?
The first is tantamount to a censorship argument. Those concerned about the ‘net's neutrality in this debate argue that network operators could, if they were commercially or politically aligned to Google, slow down or entirely block the delivery of the service of a competitor such as Yahoo or MSN. This would render Google's competitors unusable and force internet users onto Google.
It is technically possible for an ISP to create a walled garden, only allowing its subscribers access to a particular search engine. It is even possible for a transit ISP, one that carries data traffic between a search engine's hosting infrastructure and the end user's ISP, to enable faster traffic transfer to a particular search engine. They just offer a larger pipe with more bandwidth to their preferred search engine than any search competitors. They may even decide to pay for a CDN to bypass sections of the ‘net by providing a private highway from its hosting infrastructure to the ISPs with most users-thereby giving a better service to those ISP's subscribers than perhaps others not peered by the CDN.
Debate Two - Data Packets
The second debate is very different. It concerns whether the transfer of data packets-when downloading music, uploading documents, accessing web based databases, using particular VoIP systems or watching live Webcasts-should be treated with equal consideration. The question is whether there should be a hierarchy in importance of such "classes" of data.
The internet is a great leveller, and the egalitarian nature of the internet community suggests that these classes of data should not exist.
It's a wonderfully idealistic approach to networking. It is, however, unrealistic. Networks are not homogenous. Nodes on the network are not equal. Links in the network are not all the same.
The Internet is a polymorphic environment. No two networks are the same. The ‘net has come to be from merging many telecoms networks and using the same language-the Internet Protocol. As such there is no such homogeny. The only way to ensure that all users had the same experience would be to adopt a lowest common denominator approach-which would mean that search engines and other online applications would have to host their servers on a 9.6Kbps line and that all central routing infrastructures would likewise have to throttle down to 9.6Kbps. Given the amount of data flowing around in the many multi-gigabit per second links in the core of the Internet today, this would mean that it would take literally years for most emails to get through.
Obviously that's not viable.
Those in favor of a technically neutral internet try to take a "highest common denominator" approach. They say that over-provisioning of all nodes and links on the network will mean that no data has to queue and that all data will be delivered in good time.
It's not a real-world argument, though. We've seen time and again that the more capacity each network operator provides, the more data end users find to fill the pipes with, such as TV over IP.
Internet Class
There have been great strides made in preparing the internet for "Class of Service" based traffic. For example, the next generation of Internet Protocols (IPv6) use a Class of Service Header. This means that IPv6-aware routers, of which there are already many, can pick out live video data, voice data, and internet radio data and route it via specific quieter networks away from the general data that email, web, and download traffic comprises. This has enabled internet telephony, broadband TV and so on to enrich and diversify the Internet.
Internet users will always want an ISP that provides the widest access.
It is very difficult for network operators to both build their subscriber bases and wall their subscribers off from content. End users increasingly operate their own content filters and if they feel that that choice is being exercised for them they are likely to "churn away" to a provider that allows wider access. This flies in the face of those involved in the first debate who believe in a non-neutral internet as regards content. The "content is king" argument just doesn't hold any water.
As for those in second debate regarding Class of Service, a scheme that would give equal priority to every type of data package would retard the growth and value of the internet. Operators of varied and diverse networks often struggle to cost-effectively meet the demands of end users and have nothing to gain by making the internet slow, unwieldy, and inefficient. Quality of Service and Class of Service are a natural development and essential to allowing service providers and end users to make and take use of the Internet.
An internet where all data has equal priority is akin to a road where all traffic has equal priority. Imagine a road where there are no rights for ambulances to cut through, no bus or cycle lanes, cars driving on the sidewalk. Building more and more roads could be an answer, but an inefficient one.