Saturday, March 4


PBS's Robert X. Cringely crunches the numbers on video over the internet and finds that it'll take 20 years given Moore's Law to even begin to reach the bandwidth requirements to deliver real-time TV over the internet. That is given the current centralized server distribution model. The solution, P2P.

Bittorrent can scale to meet the needs of distribution already, but it doesn't stream video. Cringley goes into several potential P2P options in this amazingly informational article coming to the culminating point that ISP's though now fighting a battle for control over the data that flows over their networks will eventually grow to be big supporters of P2P distribution as ultimately increased network usage will prove a windfall.

Link: PBS | I, Cringely . March 2, 2006 - Peering into the Future

My take...

First of all I've been talking about this issue of the adoption of TV over cellular networks to anyone who will listen... So called mobile TV is at it's present state completely inept. The single greatest reason is because cellular carriers are trying to control everything that crosses their networks allowing only exclusive video content over their networks... there is therefore now very little choice and very little offerings for video over cellular networks. Very few options means very little incentive for adoption. On the other hand there's another model... allowing anyone to author content for delivery over cellular networks... even encouraging it. This will create a myriad of options for consumers and all sorts of options driving adoption. Meanwhile, much like in Cringley's scenario the carriers will make MINT based on the demand for bandwidth. This want for a hand in content has one simple reasoning, greed, and it's actually hurting them as it slows adoption.

As for cringley's article. I must respectfully point out the following points.

1) While cringley does prove that mass distribution of video over the internet is possible, he fails to compare it to the cost of delivering popular media over current broadcast mechanisms. The internet has a long, long way to go before it even begins to compete with satellite and cable in delivering populist media.

2) Therefore big TV and Cable satellite markets have no incentive to jump into delivery over the internet in mass given the much cheaper cost of broadcast delivery, even if it is in the interest of ISP's, which may be true but they don't necessarily have a clue.

However, it does make sense for the other 99% of the world that doesn't own or have access to broadcast mechanisms... it's here that innovation will happen and indeed it already is... especially on the fringes with small and midsize publishers and businesses just outside of the reach of broadcast. In talking to midsize publishers I find that such technologies as audio and video podcasting are just what they ordered, and what with over 20,000 audio podcasts and 6000 video podcasts it's hard to say otherwise.

Innovation does NOT happen from the top down. It happens from the bottom up... sometimes in the middle... but always where there's the greatest disparagement in the network. With broadcast media... it's those publishers just outside of the loop.

3) Bonus round... While I'm very interested in these other P2P technologies cringley mentions ultimately I think bittorrent and open standards have the edge. In short I believe bittorrent will be the first likely candidate for reatime streaming media. Indeed there are already people working on it though their are tremendous technical considerations, especially given the average up-speed on most home DSL and cable modem connections, but those on highspeed connections could be able to watch a video in real-time in within the year. Then all we need are video aggregators and players with integrated bittorent like the Windows version of FireANT

Technorati Tags: , , , , ,

No comments: