Blog Post

Bram Cohen: My goal is to kill off television

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

BitTorrent inventor Bram Cohen demoed his P2P live streaming protocol at the San Francisco MusicTech Summit on Monday, which he said could potentially stream live video to millions of computers with no central infrastructure. Cohen said that the protocol could potentially be used for video conferencing, live streams of video game tournaments or even live sports events. “My goal here is to kill off television,” he joked.

Cohen has worked on P2P live streaming for a number of years, and told us a while back that he completely had to start from scratch because traditional P2P algorithms introduce too much latency for live applications. BitTorrent Inc. hasn’t said how exactly it intends to productize the protocol, but Cohen said on Monday that he is talking to a number of potential partners. BitTorrent has also started to run a number of  field tests on its website in recent months, streaming weekly live music events with the P2P protocol.

The ultimate winners of a P2P-based solution could be consumers, he argued, because it would enable publishers to put much more content online at a fraction of the cost of traditional CDNs. “Most of the video that people consume today is still not on the Internet,” said Cohen, adding that existing protocols aren’t set up to support big live events.

24 Responses to “Bram Cohen: My goal is to kill off television”

  1. Guy Barsheshet

    I’ts not about live, it’s about STREAMING…
    If we can stream live content we can also stream offline content including Movies and TV…
    Now that is the real Hollywood killer.

  2. Availability of P2P technology in itself is not going to do anything. The fundamental issue is still the “content” itself. You have to own good content and the production or acquisition of good content comes at a cost as Dan Rayburn points out. You cannot create quality programs without some central infrastructure. While this may work for P2P conferencing and video game tournaments, live sports broadcasting is not a joke. You have to be able to provide multiple camera feeds, professional camera operators, with knowledge and ability to follow on field action, and also have technology for A/V post processing, dealing with rights issues (NFL, MLB, ….), Ad-insertions etc., and also have a central infrastructure to be able to provide switching between multiple viewing angles, instant replays, SAP etc etc.

    So while bold statements like “Kill of TV” make for nice headlines, it is impractical today to achieve the same level of end user experience with P2P (with no central infrastructure) as it is with existing live broadcasting of events. I can see a day in the future where P2P aided with some sort of cloud based setup for post production, mixing and delivery for a limited set of sporting events (with acceptable latency). But I would argue that mere P2P – with no central infrastructure will not “kill of TV”.

  3. Laurent Marie Coq

    Any folks that have “played” against big fishes are either dead, have disappeared, or are now living in a trailer on a Doritos/beers diet…. Solong Bram, it was nice knowing you!

  4. thelonepunman

    Here’s a thought…have the traditional broadcasters as the PLAN B when the wired/wireless system collapses when Mrs. Murphy’s great-great-great-great-grandcow kicks over the flourescent lamp in the milking parlor and the resultant mercury vapor causes a “HAZMAT CHICKEN-LITTLE DEFCON 9 LEVEL” event that causes every little Tom, Dick and Hillary have to get on their FacePlant, TwitterRista, iPuck, iFone, Crackberry device (ad nauseum) at the same time and bring about the biggest ClusterFlock DNS the world has ever experienced or seen. And while every one who thinks they matter decides that they have to dial 9-1-1 to report the event to their aunt’s mother’s soothsayer and part-time Mayan spiritialist (for you EOTWAWKI-types)…your local broadcaster will still be able to tell you that the offending cow has been PETA-certified as still being healthy, safely out of the milking parlor and pooping on some grassy knoll near a downtown Dallas Parkway. BECAUSE, if you don’t have a PLAN B, you have anarchy. BTW, does anyone remember what happens when all the eggs are in one basket? It appears the FCC still doesn’t have a clue and continues to operate that way. Go figure. Enough soap-boxing…

  5. Cool Technology. To be frank I still don’t have a clue as to what the article pointed out. Tech-dumb here. But I like the thought expressed by Bram. Hope to see something from his company hit the markets soon.

  6. Peter John

    Hasn’t a service like this existed in China for a while now? Not sure exactly how its done now, but I remember a while back I found a Chinese app for windows called PPTV that essentially used bitorrent technology to deliver the video (the more seeds, the better the quality/connection). Took a while to buffer(?), but I was getting decent streams of NatGeo on that thing.

  7. This is laughable. Does Bram Cohen read the newspapers at all? Live video streaming has been around for many years now, and works well in many parts of the world. It’s the US that is behind. In other countries, it is quite common today to watch streamed live television on PCs, tablets, mobile phones and big screens. Some of the providers use P2P technology, others use CDNs. Either way, live video streaming is now a SOLVED problem.

  8. matthewfabb

    So the question is what does this do that the latest Flash Player cannot do? As ever since Flash Player 10.1 there has been functionality for peer-to-peer video, audio and even data. It’s how Chatroulette was launched by a 17 year old, with very few server costs. Basically just the costs to launch the initial app, the video streamed from one client to another without having to go through a server. Streaming a live event via P2P in Flash Player uses something they called multicast. In that the video is sent out and branched out from client to client. I know Adobe has used it successfully to stream their big conference Adobe MAX. Not sure where else has used it, but I imagine there have been places.

      • matthewfabb is exactly right. P2P has been available in Flash Player since 10.1. And it has been used to do very large scale live shows. “multicast” that he refers to is not just native IP multicast (which doesn’t really work over the public internet). It is referring to application level multicast, which happens over the p2p mesh. But it is not just that. Where native IP multicast is available, it is a seamless blend of the 2. Where you can get the data over native IP multicast (most efficient), it will. And where it cannot, it will fill it in via p2p (or app level) multicast. To the end user, they just ask to play a stream, and it will magically get delivered by 1 of the 2 methods discussed above, or BOTH at the same time. Similarly, for the publishing client, the complexity to manage the different delivery mechanisms is hidden. It is actually quite powerful, and works very well, and the protocol has many desirable properties that BTL may not.

  9. P2P is a waste of time. The cost of bandwidth is not a problem, that’s dropping by at least 25% every year, it’s the cost to license the content that is the problem.

    Netflix spent $3B on content licensing and less than $50M on video delivery for all of last year, so delivering the bits is not the major cost for content owners. Customers biggest cost is the entire video ecosystem, not shipping bits from point A to point B.

    Every new company that comes along always says the can lower the cost, but the cost is no longer a problem. P2P vendors have been pitching this ideas for years, with zero success and not a single major content platform uses a P2P distribution platform.

    Lots of hype and talk yet no actualy adoption and a solution that’s not needed in the market.

    • George White


      While I certainly agree that content licensing is the biggest part of the cost of a typical content stream, for free to the consumer, ad-supported content, distribution costs are still a big risk, especially, for small and self-publishers (artists, for example) of potentially viral content. This is why Ustream, etc are in business. I’m interested in seeing just how inexpensive it can be. More competition should be good for consumers, and there is some disruptive potential here.

  10. That’s exactly what I’ve been dreaming of. For an independent, non-commercial broadcaster like me it would be a great way to share good quality (I hope) streams on low budget and to involve viewers into the sharing process (and go viral). I can’t wait to see this protocol in action. Where do I sign up for beta testing? ;-)