Broadcasting Online Video


Video services are drumming up fears that the Internet’s infrastructure will be overwhelmed by the mass amounts of bandwidth such large media objects use for distribution.

A common response to this growing fear is to point to the increased capacity planned by network providers. Unfortunately, this avoids a basic issue: The Internet is designed for point-to-point communication and not the mass distribution of large media objects.

The rise of Internet-distributed video has been rightly lauded for democratizing video publication. Today, anyone with a camcorder or Webcam can create a hit video and distribute it worldwide.

And as consumers demand more video on their PCs, traditional media companies are stepping in with their own robust content. Hit shows currently watched by millions of viewers over traditional TV are increasingly redistributed on the Web. To watch our favorite shows whenever we want, we no longer need to use the arcane process pioneered by TiVo, whereby content is converted to an analog signal, broadcast and captured according to time codes and saved back into a digital file. We can now just get that file directly from the publisher.

However, broadcast is far more ideal for mass-market video distribution than the Internet. Whether there are 10 viewers or 10,000, the economics are the same.

Over the Internet, it costs 1,000 times more to reach those 10,000 viewers. That is because the identical file is sent to 10,000 end points individually. Peer-to-peer (P2P) methods may hide this cost by offloading the bandwidth to the users, but the fundamentals do not change.

What few people realize is that the Internet can broadcast, with all the efficiencies that it affords, through a process called multicasting. In this process, a file is sent with no particular receiver identified, however multiple receivers can “listen” for that file and grab it as it comes by. If a cable provider identifies that hundreds of users within a given proximity are regularly downloading full high-definition copies of a television show, that provider could recoup the drain on its network by multicasting each new episode. The content provider avoids the delivery cost and the network operator is able to allocate more bandwidth for voice-over-Internet protocol and other services.

Many startup video distributors have dreams of becoming cable companies, offering thousands of high-quality video channels over the Internet with no geographical restrictions to their growth. However, underlying their business model is the fact that they are offering a competitive service over the very same pipes operated by their competition.

A better approach is to work with the cable infrastructure and build a true broadcast network for content that is by its nature designed for mass-market distribution.

The beauty of the Internet is that it can go both ways. Point-to-point distribution of video from small and independent producers should be encouraged, but broadcast is best for distributing the next episode of a widely popular TV show.

In either case, persons wishing to view videos get files on their hard drives. But the efficiencies gained through multicast make the Net better for everyone.