The (impending) arrival of Giraffic, with its Peer-2-Peer content delivery technology, raises some interesting thoughts about online video models.
Obviously, the first question to strike me was: what if YouTube was powered by a P2P delivery network? Imagine a distribution network comprising hundreds of millions of “mini”servers, serving up over a billion video views every day.
It’s a WOW factor on too many levels. I won’t even bother going into the cost-savings that would result from a scenario like this. Bandwidth costs could be halved, at least.
But even more exciting than the cost savings is how this could drive a fundamental change in new media business models. Now, you have millions of “distributors” who are not just pointing people to your content, but actually serving up the content as well. As a content owner, you benefit. As a content consumer, you benefit.
As a CDN, you’re scared stiff, and as a video platform, you’re ensuring QoS and reliability, perhaps giving the CDNs a comforting pat in the back.
- – -
A few harsh realities that stand in the way, nonetheless.
P2P delivery happens in chunks, so for users to help deliver those chunks, data will need to be downloaded/cached on their local machines first. This is obviously antithetic to streaming video, and a big argument in favor of progressive downloading.
Along the same lines, the local P2P client would be the Flash player. Every time a user refreshes a page/navigates away from a page, the P2P network loses a seed. This is in stark contrast to a traditional BitTorrent client, which is “always on”.
- – -
Nonetheless, P2P delivery of online video is a very interesting conversation to have. Adobe is definitely thinking hard about it, with RTMFP (Real Time Media Flow Protocol), supported by the Flash 10 Player. Joost gave it a shot, but failed.
I think this is one disruption that will completely depend on timing, more than execution. Now’s not the time for it, but 2010…who knows?