Business Model Premise: The current Internet Service Providers and CDNs at the core can not effectively handle the video load across a single or even multiple locations:

a)Backbone, peering interconnects, and the hundreds of thousands of routers deployed can not handle the load or performance (latency, jitter, etc.) for video.

b)The leaf nodes (i.e. Grandma’s PC) in aggregate have the cycles and network capacity, if shared, to distribute popular video content today. Jeffrey Payne (Grid Networks) claims that less than 1% of the aggregate last mile content is utilized today.

c)Popular content can be chopped up into small chunks such that many downloaders become sources, and topologically close downloaders will prefer the topologically close sources. This ‘swarmcasting’ requires only a source ‘seed’, and a lookup mechanism for the first downloaders to find the seed, and then to direct future downloaders to topologically closer sources. This has proven a viable technology even if the business models have been flawed in the past.

Peer-to-peer is compelling to the VSP since the load on the VSP infrastructure and therefore cost is minimal. With this low cost of entry, it should not be surprising that the author has discovered the most innovation occurring in the peer-2-peer space.

Peer-to-peer does however count on the good graces of the last mile operators allowing their customers to fill their pipes with this shared video content.

Cringeley suggests a Utopian peer-to-peer approach whereby 2.5% of the “Desperate Housewives” audience of 10,000,000 distributed randomly across the net could be given a feature like fast forward over commercials if they allow their machine and DSL line to be used in a peer-2-peer reflector mode. Then the rest of the 10,000,000 households would pull the pieces of the video files from these randomly distributed reflectors.

This spreading of the load is done with the expense of traffic distribution spread across the interested user base network infrastructure. The cost to the content provider here is only the cost of distributing the content to the 256,000 reflectors and perhaps some control plane overhead to manage this.

There will always be at least an initial spike in transit fees while the “seed” content is being distributed to “peers” in the peer-to-peer desktop sense. The transit load continues until the content is distributed to desktops topologically closer to the end-users requesting the file. Many p2p users today move completed downloads out of their shared directories making their sources unavailable to others. In response, many peer-2-peer networks use a give-to-get throttling tracker mechanism to encourage those people to keep their PCs seeding the content. As long as some desktops remain online and the source remains available, peer-2-peer remains a powerful and inexpensive alternative for distributing large scale video traffic.