Network Interruption

Internap is a company that toyed with the CDN market and had its fingers badly burnt. Since writing down its investment in Vital Stream and focusing on its core market, it seems to have gone from strength to strength.

There is no market stranger than the CDN market. It is the only truly scalable part of the Internet TV market since neither channels nor VMSes seem to be worth anything, but it's horribly commoditised and the major players (Akamai, Limelight) could be easily usurped shortly by their main suppliers (Level3, AT&T).

There are only a handful of scale video customers - maybe a hundred worldwide worth more than 100TB pa in my estimation, so this is a highly competitive market (most revenues are made from web or upgrade delivery, not video).

The smaller players (e.g. Mirror Image, Edgecast, Velocix and Panther Express) cannot cope with the bigger contracts, which demand bursts that require very large networks which are, in effect, nothing to do with their day-to-day business, and are better catered for by cloud networks (if such things really exist).

Meanwhile customers are still struggling with the failure of P2P (which is not absolute yet), the insatiable demand for unicast streaming and the lack of advertising to pay for this (actually, there's plenty of advertising bucks, it's just that audiences are being delivered in too small quantities to be of interested to ad agencies brought up on the fictionality of TV viewing figures.

Comments

phoneranger said…
Failure of P2P? Well if you mean Joost, RawFlow, BitTorrent etc. those were failures of business models not the technology. Using P2P to deliver fragmented longtail content on demand has been a failure for sure. In fact, P2P's big weakness is that is doesn't "scale down". But as the YT live event showed, there is an audience for "fathead" content. And it's dubious that CDNs can scale enough without help from P2P or P4P or P6P one day.
Iolo Jones said…
Agreed on the longtail, but I'm not sure what the YT live event was. However, I've yet to see a) a successful P2P service or b) a P2P model that actually saves money for its customer.

In the UK Knotiki was dropped faster than lead on a balloon after the idiots who thought that downloads was the way to go were thoroughly disproved by their audience and had to turn to streaming, where P2P has minimal benefit.