Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > Networks

Internet overload? No, go beyond P2P

Posted: 24 Sep 2007 ?? ?Print Version ?Bookmark and Share

Keywords:Internet overload? blogs? bandwidth shortage?

John Riske

EMarketer reports that online video viewership is projected to expand from 62.8 percent of Internet users in 2006 (114 million viewers) to 86.6 percent by 2011 (183 million viewers), and iSuppli recently reported that consumer appetites are migrating from clips to long form content, such as sports events.

Chicken Little has evidently logged onto the Internet, and his warning that "the sky is falling" has taken on a modern tone that is echoed in blogs and other news sources: the Internet will collapse under the burden of so many viewers watching so much video online.

Myth vs fact
One myth to dispel right away is that the world is going to run out of bandwidth. The telecom industry overbuilt capacity during the last boom, but, as that is filled, there will be opportunities to make more money on IP capacity (whether fiber to the home, WiMAX or other spectra), and someone will step up to meet the demand.

The real question is how to ensure a high-quality viewing experience on the Internet while leveraging resources in an economical way, so as not to "waste" the available bandwidth. I'll address the shortcomings of video delivery as it is today and explain why decentralized peer-to-peer (P2P) networks are an inadequate solution, and I'll suggest a few directions to explore as we collectively forge the future of video delivery on the Internet.

Unlike cable, the Internet is a "best effort" network, meaning that there is no guarantee a piece of data will travel between Points A and B in a given amount of time. The Transmission Control Protocol allows two points to communicate with one another to determine which bits need to be sent again, but the multiplicative effect of latency and packet loss means that, on any broadband connection, there is a limit to achievable throughput.

Tap P2P nets?
Content delivery networks put servers in close geographical proximity to end users to reduce latency, but video is notoriously sensitive to latency and packet loss. Point-to-point connections are fragile, and a single glitch along that connection can disrupt the delivery, causing video playback to suffer.

Decentralized P2P networks make connections to multiple sources of a video at the same time, doing away with a transfer that can suffer from a single point of failure. These networks have become very popular. In fact, there are estimates that as much as 40 percent of global IP is traffic generated by peer-to-peer users.

P2P networks get a couple of things right: depending on many sources for a video is better than depending on one. If you're looking to distribute a video for free, P2P networks might be for you. However, delivering video for free is a benefit only for the entity footing the delivery bill?not a benefit for the end user. When it comes to user experience, P2P networks don't deliver.

Consumers are used to the experience provided by traditional platforms such as cable and satellite. Most P2P networks make a user wait until the entire file is downloaded and assembled before viewing. Moreover, there is still no way to ensure that a piece of content is delivered in a timely manner. This is true because P2P networks rely on users to contribute pieces of a file to other end users, and users' upstream bandwidth is largely limited by their ISPs; thus the overall bandwidth contribution in a P2P network is limited.

Move beyond P2P
The subject of ISPs and P2P networks has been in the news a lot recently, with Comcast throttling peer-to-peer traffic, blocking uploads and further slowing downloads. The reasons for this are myriad. One reason relevant to this discussion is that P2P traffic is not cacheable. This means that every time a piece of a file is requested, it hops across various networks to its destination. This is a burden on ISPs, which have to pay for peering with other networks.

The bandwidth burden created by non-cacheable protocols is not limited to P2P networks. Any nonstandard protocol wastes bandwidth?Windows MMS streams are not cacheable either. If four people are watching an MMS stream at 2 Mbps, an ISP is serving up 8 Mbps of that video. Instead, video should be delivered using cacheable protocols such as HTTP, taking advantage of the billions of dollars invested in HTTP proxies, caches and servers that are the very fabric of the Internet. To utilize resources efficiently, the use of standard, cacheable protocols is critical.

To bring the discussion back to user experience, a good lesson from P2P networks is that the more sources of a file there are, the better. After years of expanding its infrastructure, much of the potential of the Internet is latent in the form of unused HTTP resources. It's time to get creative and start tapping these unused resources, using standard protocols to effectively and efficiently deliver online video. It's time: get beyond a fixation on peer-to-peer networks.

About the author
John Riske
is business development manager at Swarmcast.

Article Comments - Internet overload? No, go beyond P2P
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top