Even if growth of the Internet was huge, the arrrival of
video was a revolution. And yet, people are still
looking for a killer app to justify NGN build-out. I think it is here now, and it is called streaming video.
Up to now, we mostly live in a download world. No matter if it's a web page, a .pdf file, an MP3 file or a movie, we basically download (causing a traffic burst) and then read/listen/watch. Even YouTube fundamentally is a download service. And after downloading, we are pretty much offline for a while.
This ties into a FTTH project that I had a chance to talk about with the operator. It has roughly 100 homes subscribing, with services ranging between 20 and 100 Mbps. But here comes the shocker: the aggregated bandwidth used never crosses the 100 Mbps mark!
Sandvine recently published its
2009 Global Broadband Phenomena report, claiming that streaming video is exploding already (to 27% of total traffic, from 13% in 2008). However, they include YouTube in the 'real-time entertainment traffic', while fundamentally it is a download service. Further, the top 1% of subscribers account for 25% of traffic. I would add that this 1% is not a constant group, but that all of us are part of it at different times.
Obviously, in a streaming world, traffic will explode. Imagine the continuous flow of 10 Mbps caused by watching HD video, as opposed to the bursts from downloading. The consequences will be especially large for shared networks (cable, PON, wirelsss). If
Nielsen's law is to hold, it looks like dedicated fiber is the only infrastructure that will keep up.
So here is the point I would make, based on the above:
In a download world, any NGN will do, but in a streaming world we need dedicated fiber.