With Joost's launch upon us, BitTorrent going mainstream, Akamai buying Red Swoosh and a raft of other peer-to-peer (P2P) initiatives underway, it's time to consider legitimate P2P's possibilities and pitfalls.
First a disclaimer: I don't pretend to know all of the technical ins-and-outs of P2P, but I think I know enough to be dangerous. Here's my current take: P2P has a ton of potential as a legitimate distribution platform, but has to navigate some significant challenges if it is to succeed.
A P2P Primer
For those of you new to the P2P game, in essence, P2P's big advantage is that it allows users themselves to become servers of content to other users. In doing so, the load for delivering content is shifted from central servers to the "nodes" or users on the P2P network. Until relatively recently, P2P was popularly associated with the illegal "file sharing" networks (Napster, KaZaa, etc.), most of which were (and still are) used by users to swap audio or video files without permission of the copyright holder. Users could look up where certain content resided and then download it accordingly.
What's new about P2P is that many (e.g. Joost, BitTorrent, others) see it as an important, if not essential, way for video to be legitimately distributed. P2P companies argue that the Internet's current architecture cannot effectively scale to deliver large quantities of video (especially live streams) in an economic manner. Since P2P gives users the ability to directly share with other users, P2P also has a potentially disruptive effect on the overall value chain and how video aggregators continue to establish value for themselves. P2P requires users to install client software on their computers. These clients are then available on the P2P network, sending files to subsequent users requesting content that they have already stored. In the case of video or audio, files can be delivered for either download or streaming.
All of this is intended to happen invisibly to the average broadband video user. Of course, to nobody's surprise, the average user couldn't care less how video actually gets to his or her computer, as long as it gets there quickly and in reasonably good shape.
P2P is a potentially big deal for the biggest broadband video content providers. That's because delivering large volumes of video in the traditional client-server paradigm is still pretty expensive, notwithstanding the significant declines in content delivery networks' (CDN) pricing. With everyone forecasting huge increases in broadband video consumption, together with larger video files (due to better encoding, High Definition, etc.), getting a handle on delivery costs is a key challenge for content providers.
Compounding matters is that broadband video business models remain relatively immature, so expense containment is all the more important. P2P allows these content providers to shift all or some of the responsibility for video distribution to the users themselves, while establishing direct connections with users (i.e. no 3rd party distribution costs). The users' computers are leveraged for both storage and delivery, while the bandwidth is essentially free, since users upload content using the local broadband ISP's network, not the content provider's CDN service. If P2P succeeds, its potential to cut content providers' delivery costs, while delivering high-quality video, is obviously very significant.
Important Challenges Lie Ahead
Of course, potential is one thing, reality is another. From my vantage point, consumers' willingness to become P2P nodes and ISPs' restraint in blocking P2P traffic represent the biggest obstacles to P2P's future success. First the consumer acceptance challenge. Getting the P2P client on millions of users' computers or into their living rooms is not trivial. In this era of spyware, malware, viruses and other technical nuisances, mainstream Internet users are becoming more reluctant than ever about loading anything onto their machines that doesn't come from a recognized and trusted brand. Since P2P's whole promise relies on files being propagated to many users, anything that limits this from happening is obviously very detrimental to P2P's success.
Then there is the even thornier issue of how broadband ISPs are going to react to users clogging up precious upstream bandwidth by serving as nodes. Virtually all American broadband ISPs offer "asymmetric" Internet access, meaning that the amount of bandwidth offered in the upstream path is usually only a fraction of that provisioned for the downstream path (this is due to some fundamental limitations related to the way that ISPs' networks are allocated). Re-architecting these networks for potentially burgeoning upstream traffic flows would be cost-prohibitive and a non-starter.
To date, broadband ISPs have used "traffic shaping" technology to identify and limit P2P traffic. They have also kicked customers off their networks who have used too much bandwidth (a little secret in the industry). All of this has been sort of OK to do when most P2P use was for illegitimate file sharing. But what happens when it's for legitimate use, such as Joost or the newly legitimate BitTorrent? Limiting users' access to their full broadband service is going to evoke howls of protest.
And of course, remember that the net neutrality proponents are waiting to pounce on any sign of broadband ISPs de-prioritizing or worse, blocking, certain types of traffic. Net, net, a big wildcard in P2P's success is how ISPs are going to react.
Planning for P2P Success
P2P proponents need a game plan to overcome these looming issues. Here's what I think makes sense: Well-established branded content players will need to take on the primary role for P2P client distribution. Of course, this approach has been used for previous media players' distribution (i.e. Real, WMP, Flash, etc.) and for updates. We've all had the experience of being asked to download player software or an updated version of previously installed software. P2P client distribution could be no different.
But what will incent major content providers to assume this responsibility on a mass scale? They'll have to see real (not theoretical) business cases for delivery cost reductions and quality improvement. Of course, getting paid to become P2P client distributors (either in cash, or as part of distribution deal discounts, or some hybrid of the two) would also clear the way. Companies like Joost and BitTorrent need to remember that while their brand awareness among the Internet's cognoscenti is high, among more mainstream users it is still low. So leveraging their content partners' brands to turbo- charge distribution is key.
BitTorrent, for one, is already doing this with their BitTorrent DNA
technology. Another opportunity for P2P client distribution is embedding it in various consumer devices. For example, BitTorrent also offers a software development kit (SDK) that consumer electronics and chip makers can use to embed the P2P client in devices. This removes P2P download complexities for users, and is intended to make P2P usage completely invisible. The ISP solution seems more complex.
Some believe that ISPs should look at P2P as a business opportunity to deliver a quality-of-service (QOS)-guaranteed platform to the P2P application providers such as Joost and BitTorrent. This would be accomplished by installing caching servers in broadband ISPs' facilities. These would essentially allow ISPs to serve content locally, mainly relying on the P2P protocols to deliver from the caches when appropriate, instead of from the nodes. This approach would preserve upstream bandwidth and limit ISPs' need to increase their peering capabilities to handle video coming in from the Internet backbone, while also leveraging P2P's scalability.
This "peer-assisted" approach may be the optimal migration path to P2P adoption from an ISP perspective. Though the economics still need to be fully fleshed out, I've heard a pretty persuasive argument for this model from a company named PeerApp
(disclaimer, they're a client), which is worth understanding further if P2P affects your business. One way or another, ISPs need to be brought into the P2P fold. Simply ignoring them or relying on their reluctance to tempt the net neutrality gods is not a sound business approach.
P2P offers very exciting potential to enhance users' broadband video experiences. For content providers, it holds the promise of profitably scaling up their broadband video activities. It will be very interesting to see how key P2P players navigate impending challenges to their success.