August 2008 Archives

After the mad rush to get on the video bandwagon, it seems that enterprise IT executives are waking up to the cost of moving all those bits.

Steve Vonder Haar over at Streaming Media magazine explains the problem: Eyes on the Enterprise: Mo’ Videos, Mo’ Money—Solving the Problem of Network Cost.

What most of the IT execs surveyed miss is the solution. They’re putting their faith in network expansion, wider use of CDNs and P2P apps.

To quote Steve: “Against this backdrop, it can be maddening to try to understand what some IT executives are thinking when considering the deployment of networking solutions to address multimedia distribution issues within their organization. The bulk of respondents in the IMS survey (68%) said they would prefer to address the need for expanded multimedia distribution by increasing the capacity of their existing corporate networks. Essentially, in a multimedia networking market where costs are king, IT executives are saying they prefer the most expensive, Cadillac-style solution to their troubles.”

Maddening, indeed, when they could reduce their costs immediately with better video compression while not requiring any additional network capacity.

For example, an enterprise or UGC site using Adobe Flash Player video encoded in H.263 could shrink their bandwidth use by at least 25% (in some cases up to 40%) by switching to On2 VP6 or H.264 video in Flash.

I received a white paper last month called “10 Ways to Go Green with Video.” The paper is pretty good, and cites many of the familiar arguments about video in the Green economy—namely that it can replace carbon-intensive activities like plane travel, automobile commuting, on-site training, and so on.

These are valid arguments, of course, but there’s a dirty (no pun) little secret in this kind of thinking. Any video application that you deploy requires electricity (not only to create, edit & view the video, but to send the video packets through many routers to span the networks).

Unfortunately, most electricity is still generated using coal and other carbon-intensive fossil fuels. In 2004, coal-fired electricity generation accounted for 41% of world electricity supply; by 2030, its share is projected to be 45% U.S. Energy Information Administration (PDF).

That’s a 4% increase over 26 years, which doesn’t sound like a lot. But when you consider that the projected consumption of electricity worldwide is expected to grow from 16,000 to over 30,000 billion kilowatt hours by 2030, that translates to a lot of coal. In fact, Google is so concerned about future energy costs & sources that they’ve proposed hosting their data centers on ocean barges to harness wave energy.

I’m not a carbon scientist, so I can’t say whether the carbon emitted to make the additional electricity required for increased global use of video is more, less, or equal to that generated by airplane travel, automobiles, etc.

But I do know a thing or two about video, so I’d like to suggest an 11th way to “Go Green with Video”: Use Better Video Compression.

At its core, video compression is about efficiency. Yes, better compression results in better video quality, but it achieves quality through smarter allocation of available data in the video stream.

But an aspect of compression that doesn’t get mentioned often is this: Good compression also reduces the amount of computing power required to play back (decode) the video. For example, our new On2 VP8 codec uses up to 50% fewer processor cycles to decode than other advanced video formats. This means less electricity to play video, less electricity to play higher resolutions, and more hours of video playback between battery recharges on mobile devices.

This isn’t to say that On2 Video will singlehandedly solve the world’s energy problems, but on such a massive scale as this, every little bit adds up.

A 12th way to go green with video is this: Do it in hardware. But that’s a topic for another day.

A recent white paper from Cisco has everybody talking again about the role of video in the coming IP data “exaflood.”

In the paper, Cisco predicts that worldwide IP traffic will surpass half a zettabyte (522 exabytes to be exact) by 2012. Previous reports on the same theme have been published by The Discovery Institute and IDC.

Cisco also estimates that, “The sum of all forms of video (TV, VoD, Internet, and P2P) will account for close to 90 percent of consumer traffic by 2012. Internet video alone will account for nearly 50 percent of all consumer Internet traffic in 2012.”

Think about that: 470 of the 522 exabytes of data traversing the tubes in 2012 will be video.

Whether or not the Internet can handle this volume of data is still under debate. But what if even marginally better video compression could decrease that figure by, say, 20%?

That would translate to a savings of over 94 exabytes per year. That’s 100,931,731,456 gigabytes.

For argument’s sake, let’s assume that all bandwidth in 2012 will be sold at one very low price: $0.05 per gigabyte.

Our theoretical 20% reduction in bandwidth would mean a savings of $5,046,586,572 per year.

To anybody except the U.S. government, that’s a lot of money. And of course the likelihood of all bandwidth, everywhere, costing $0.05/GB is slim.

Consider also that NHK Japan has been hard at work since 2003 developing Super Hi-Vision (UHDV), a video format that uses 16 times more data than today’s 1080p HD. Their goal is to begin broadcasting UHDV in 2015.

Constant innovation in compression will be essential to manage the explosion in IP video. At On2 Technologies, our proprietary technology empowers us to develop codecs that have consistently stayed ahead of the standard formats. For example, our next format, On2 VP8, will offer better quality than H.264 and VC-1 using as much as 50% less data.