A few weeks ago I said (following many others before, including Vint Cerf) that the Internet as we know it is not ready for live, standard-definition television. On a related note, Om Malik has recently been speculating that Google may be on the path to building just what is needed: a next-generation, TV-suitable Internet.
Let's explore what's involved in some detail.
The Internet's consumer proposition
Today's Internet (not to be confused with the web, which it enables), more than a network or a protocol, can be seen as an institution: a multitude of contracts between consumers, ISPs, CDNs, wholesale transit providers, and content providers, all of which together manage to provide a simple consumer proposition, which reads:
- If you sign up with a broadband provider you will have unmetered, always-on access to "the Internet".
- Having access to "the internet" means that you can connect to any computer (or "host") in the world that is also connected.
It is these two pillars that form the basis on which all of today's Internet applications are built: the web, FTP, voice-over-IP, instant messaging, etc. While some have claimed that a recent ruling by the FCC puts this at risk, I've argued that this is not so.
The Internet connectivity economy
Things are much more complex for the middlemen.
To allow you to connect to the wider world, ISPs need to purchase "transit" from wholesalers; this can be unmetered or not, but in the end ISPs end up paying for every bit they use.
Wholesalers in turn lease dedicated, high-capacity links to take transmit data between geographically distant points. These links --essentially cables-- could otherwise be used for all sorts of other purposes far outside the Internet.
Because bandwidth is not free, players --ISPs, wholesalers and large content publishers-- try to secure "peering" agreements between each other, whereby they meet at an "exchange" (a physical location where lots of networks meet) and interconnect for free. But to play that game you need to have as much to offer as you have to gain: others will only want to peer with you if they know that their customers want to connect to yours, because only then would they be saving money by not having a transit provider between them and you.
Not all wholesalers are made equal. Some use faster lines than others, some have more redundancy, and some have more connections. Being well connected is important, because otherwise your transit will need to go through many intermediaries (many "hops") before reaching its destination, resulting in a slow user experience.
Finally, because all of this is messy and the resulting user experience somewhat unreliable, some content providers rely on content delivery networks to largely bypass the Internet and deliver their content (web pages, clips and live streams) direct to ISPs, so that content can reach to consumers with almost no "hoops" and minimal delay.
An impressive feat
Complex as the connectivity economy is, collectively it manages to provide a simple consumer proposition: umetered access to any host on the internet.
It is remarkable that such an amount of coordination can be achieved in a decentralised, market-led way. But before we gasp in awe at the invisible hand of capitalism, let us remember how the Internet started: it was not built by the market but rather by the American government, and later by the worldwide academic community. When the private sector was allowed in, the main consumer experiences and expectations (including the web) already existed, and thus the task of building the collective contract was already largely done. Government played a bootstrap role.
Now that the Internet is in place, the fact that it stays that way is not impressive at all: any player whose product failed to enable the central consumer proposition would have a hard time making a profit. This is not to to say that today's coordination is easy; it only to say that it is no more unusual than the institution of marriage, trading or many other cases of collective intentionality.
Back to Internet TV
As I've argued before, today's Internet is not suited to live, standard-quality television, partly because the contracts in place make little provision for sustaining uninterrupted bandwidth between consumers and content providers. (I've also argued that over time this will become less and less important, as overall network quality will make such agreements less necessary. For the time being, however, without the necessary quality-of-service provisions, current Internet contracts cannot deliver a sofa-friendly, live-TV experience.)
Changing this would entail an amount of coordination similar to what was needed to build the Internet in the first place. At the very least it would entail setting up a multicast infrastructure (which is largely a matter of agreements), and adding quality-of-service provisions to all the contracts in the internet industry--a very difficult thing to implement even with the best of wills.
I doubt that the market can deliver this. While such a change would be clearly beneficial to consumers, and arguably to many other players, the fact remains that, for as long as this new Internet is only speculation, for most players it is rational not to invest in it, and for many it is rational to offer similar services (such as IPTV) in a closed, non-Internet basis.
For this to change without goverment intervention, consumers would need to demand something different. But first "something different" would need to exist--an open Inernet 2.0 where live TV could be watched, to which consumers would demand access. This would require content, destinations known and discussed by consumers, consumer habits, etc. Further, there would need to be enough competition among access providers to force all of them to offer access to this new thing.
Is GoogleNet an attempt to address both of these prerequisites at once, within the private sector, using a free proposition to create consumer habits and expectations quickly? The possibility is intriguing. With several billion to spend and Internet creator Vint Cerf on board, if Google can't pull this off then nobody can. But then, for all we know, this might not be what they are trying to do.
Finally, a reading of some of the FCC's recent decisions and statements. As I argued before, I think that while the FCC genuinely accepts that access to the Internet as we know it (Internet 1.0) is a utility, and won't be sympathetic to operators who won't treat it as such, it is (for now) refusing to say the same about the future Internet 2.0. However, its decision to limit competition in the ISP space makes a market-led breakthrough less likely than it was before.
Update (July '06): As the net-neutrality debate heats up, the DPS Projet group proposes to formalise practices like the ones I listed above through legislation that would define such institutions as
Internet.- The term "Internet" means the worldwide, publicly accessible system of interconnected computer networks that transmit data by packet switching using the standard Internet Protocol (IP)
Internet access.- The term "Internet access" means a service that enables users to transmit and receive transmissions of data using the Internet protocol in a manner that is agnostic to the nature, source or destination of the transmission of any packet. Such IP transmissions may include information, text, sounds, images and other content such as messaging and electronic mail.
Comments