Distributed Computing Solves Video Streaming Problem

March 30, 2006

Recently I learned about Network Foundation Technologies (NFT), a company providing an Internet broadcasting platform based on distributed computing. This�is an important story to tell, as many observers think�peer-to-peer technology looks like the best answer to the problem of Internet-based video streaming (see my previous blog entry, "P2P Enables Large-File Media Distribution," from March 16, 2006).

NFT's model depends on a distributed network of users who receive a video transmission as their computers and Internet connections are simultaneously being used as repeaters�to distribute the stream to yet�other users. Mike O'Neal and Marcus Morton are co-founders of NFT.�O'Neal�is chair of computer sciences at Louisiana Tech and Morton is a former vice president at EMI and the producer of several feature films.

In a recent interview, I asked them how NFT's technology works, how it solves the problem of Internet video distribution, and what this kind of technology will mean for the future of video and advertising. Here are their responses in Q&A format:

How do you see the future of online video? What will things look like in a few years?

(Morton)

Explosive growth!

While we believe that download and play "on demand" type content will remain popular, concerns over content piracy will continue to limit the amount of content that is available in this format, at least legally.�

Additionally, people generally enjoy being programmed to -- the continuing popularity of radio, despite MP3 players and CDs, is clear proof of this fact.� To cut through the clutter, we go to trusted sources and depend on them to do the work of finding and presenting the content we enjoy.� Heck, that's what a DJ in a club is paid to do -- program the mood of the club.�

People also want to feel "connected" to other people and to what is happening "right now."� There is value in being able to watch an event "as it happens," and this feature is essential for some types of broadcasts, such as sports, breaking news, weather, etc.

For these reasons, we believe that the Internet is about to experience a veritable explosion in the amount of 24-7 television-style broadcasting that takes place on the Web.� Online events, such as AOL's Live 8, CBS's March Madness, and Yahoo's broadcasting of recent NASA events, prove there is pent up demand for this type of broadcasting.�

Why is video streaming such a challenge right now? What are the limiting factors? (You could speak to the CBS example if you like.)

(O'Neal)
�
The traditional approach to streaming, where each viewer is watching a separate copy of the stream with all streams being generated by the broadcaster, simply does not effectively scale to television size audiences.� Right now, online broadcasting costs scale linearly with the number of viewers (i.e., the more viewers you have, the more you have to spend on bandwidth and servers).�

(Morton)

Yes, that's right.� It all boils down to cost.

Can you imagine a network like CBS promoting a television broadcast and then limiting how many people can tune in to watch?� Of course not!�� The more viewers a network gets, the more money they can make off of advertiser revenue, and since larger audiences do not cause their broadcast costs to go up, their profits increase.�

But in the online world, things are different.� Rationing access to the broadcast is exactly what CBS has been forced to do with March Madness.� Peak audience levels have been limited to just over 1/4 million viewers, and those viewers are "rotated," so after a fixed amount of time, a viewer will be bumped and forced to get back in line to gain access to the broadcast again.� According to a CBS press release, at one point there were 150,000 people waiting in line to watch the broadcast.� Think about the advertising dollars that were waiting at CBS’ doorstep and simply couldn’t come in the door.

Why is distributed computing the answer? Why is the NFT model working, while the CBS model isn't?

(O'Neal)

NFT can turn the entire online broadcasting proposition on its head, leading to a situation that is much closer to television and cable, where broadcast costs are relatively fixed, regardless of the number of viewers.�

(Morton)

I don’t think anybody can deny the enormous demand for online video content.� CBS’ broadcasting of March Madness was a significant milestone for our industry, as it vividly demonstrated the demand for online access to TV-style broadcasts.� The real question is: where do we go from here?�� CBS was forced to limit access to their broadcast because the underlying technology they were using simply doesn’t scale to the level of millions of simultaneous viewers.� NFT’s technology can open the flood gates for content providers.� NFT technology solves CBS’ core problem – allowing virtually unlimited audience sizes while reducing costs.� If CBS utilizes our technology next year, no one will be talking about how many tens or hundreds of thousands of people were waiting in line to access the broadcast.� Instead, the focus will be on how broadcasters and advertisers are utilizing NFT technology to maximize exposure.

Please explain NFT's platform and how it works.

(O'Neal)

NFT is a distributed online broadcast technology.� What this means is that the computers of the individuals who are watching a broadcast help deliver that broadcast to other members of the audience.� NFT’s client software is light weight, and the application is VERY respectful of the end-user's bandwidth.� NFT only uses “idle” upstream bandwidth on the viewer's computer that would otherwise go unused. The moment an end-user needs that bandwidth the NFT application immediately “steps aside.”�

Thus, from the end-user's point of view, the fact that in addition to displaying a video feed, NFT-TV players also pass a copy of the feed on to at most, two other viewers seamlessly and invisible (except for a little icon we provide on the player to make sure end-users know the status of their player).

By transferring the cost of delivering a broadcast onto the computers and Internet connections of those individuals interested in watching that broadcast, advertiser-supported delivery of 24-7 television style content becomes practical.� To the end-user, the service appears "free" because they have already paid for the bandwidth (whether they use it or not) in their monthly cable/DSL bill.�

Additionally, by properly segmenting the distributed network and directing a viewer's computer to connect to other computers within the same ISP, NFT could minimize expensive inter-network traffic and thereby actually REDUCE the costs of the Internet Service Providers.� NFT truly provides a win-win-win situation for broadcasters, consumers and ISPs.

What does NFT provide, and where is your company heading?

(Morton)

We see NFT empowering a revolution in online broadcasting.� Our company is focused on two primary markets -- professional broadcasters (via our NFTPro product) and individual home broadcasters (via our PeopleCast product).� For professional broadcasters, we will dramatically reduce their costs, allowing them to substantially increase the amount and quality of live streaming that takes place online.� NFT will also give home broadcasters the ability to reach large online audiences for the very first time with nothing more than a PC, web cam, microphone and broadband Internet connection.�
What got you into this business?

(O'Neal)

Back in late 2000 or early 2001, I was discussing the emerging Webcam phenomenon with a friend over lunch, and the question of whether it would be technologically possible to enable a home broadcaster to reach a large audience with just a computer, microphone, Webcam and broadband Internet connection came up.�

It was immediately apparent that the only way to accomplish such a task was through some kind of distributed network -- where the broadcaster would not be required to directly feed the signal to all of the viewers.� Of course, the problem that leapt to mind was, how do you maintain such a network, as viewers, and hence repeaters, continually join and depart the network.

Being a natural "egg head," I found that I just couldn't let this problem go.� Three months later, I'd worked out the core algorithms for constructing a stable distributed broadcast network that overcame the major problem faced by such networks:� maintaining stream quality given the turn over inherent in such distributed networks.�

What made our approach so extraordinary is that our algorithms view the natural turnover in the network not as a problem to be overcome, but as the engine for driving the most reliable rebroadcast nodes into the most critical regions of the network -- greatly increasing network stability, essentially "for free."��

In fact, the algorithms are so darn nifty that that is what we decided to call the company – NFT, which we pronounce, "nifty."

(Morton)

What did it for me was that I saw that Mike's approach could lead to a way for content producers to reach large audiences in new ways -- bypassing the existing "gate keepers."

As an independent producer, this appealed to me.� My first major motion picture, "Three Strikes," was produced independently.� Though the picture was an overwhelming financial success (grossing in excess of four times its development and distribution costs), getting that picture into nationwide release was a major hassle.

Many newly formed broadcast channels face the same problem I did -- they have content and know the audience is out there, but the barriers to entry on cable and satellite broadcast systems are simply too high.� NFT provides a way for these companies to directly reach consumers in a cost effective manner.

I'm also intrigued by the concept that NFT can empower a whole new wave of end-consumer generated content.� There is a lot of talent out there.� Most everyone has a message.� Blogging has allowed people to express themselves in written words, podcasting with the spoken word, vlogs with short video clips, and now NFT will bring most anyone with the desire to do so the ability to launch their own Internet TV channel.

What does the NFT model offer for advertisers? How will advertising work in the future that you envision?

(Morton)

An innovative feature of the NFT-TV player is an unobtrusive, "one click" buy opportunity.�

People are tired of flashing banner ads that distract them from the content they are interested in viewing.� Thus, the approach we adopted with our NFT-TV player is to include an informational scroll at the bottom of player, immediately below the video window.� This scroll generally contains program title and related information.� Clicking on this scrollable area launches a Web browser and points it toward a page specific to the onscreen content.� Such a page could contain further information about the program, be a strict buy opportunity, or something in the middle like an Amazon.com page that contains interesting information about the program, such as reviews, together with a mechanism to buy a copy of the video.

We envision that in many applications, the content being broadcast will, in a sense, be an ad for itself.

Thus if you are watching, “Gone with the Wind,” the scroll at the bottom of the screen would link you to a page where you could buy a copy of the movie.� Similarly, for advertiser-supported broadcast content, while an ad is playing, the informational scroll would allow you to "one click" over to a Web page to purchase the item.

We believe that the ability to give the consumer an immediate buy opportunity will be a very powerful tool in converting ad impressions into sales for certain types of products.

Overall, the NFT player offers content providers a much broader advertising inventory than ever before at a reasonable price.�� As we all know, metrics drives the advertising world.� Our player is capable of monitoring total viewership data tracking and click throughs. When developing NFT’s technology, we had the advertiser in mind because, at a lower productions cost than conventional streaming technology, any advertiser can flourish when correctly targeting relevant content.

If you could sit down with a top ad agency executive, what would you tell that person to start doing right now to get ready for the future media space?

(Marcus)

Every advertising and media buying firm is looking for new and creative ways to get their clients’ brands in front of possible buyers.� They are also looking for ways to validate and quantify their buys and budgets. NFT opens the possibilities of new ad inventory over the Internet in the form of narrow interest “streaming channels,” where audience response is trackable and data reports that can be viewed “live.”� For example, say there is a BobTV in Denver, Colo.� Bob launches his streaming channel using NFT technology and immediately attracts an average daily audience of 100 distinct viewers.� Commercials can be inserted into Bob’s broadcast a very reasonable cost…a fraction of what a network broadcast spot costs. Now, imagine hundreds and thousands of BobTV’s…Without sounding too dramatic, our technology can make it possible for very small broadcasters to profit from advertising while allowing advertisers to reach very specific audiences at low cost.�

What's the most exciting thing about the business you're in?

(Morton)

The fact that we are on the cusp of another revolution in the way people will use the Internet to communicate – personal Internet television channels – and our company has a powerful technological solution that can help bring about this revolution.

Anything else you would like to add?

(Morton)

Just that NFT is in the market for qualified investors, so if you are one and you'd like to learn more, feel free to drop us an email at [email protected].

(End of interview)

AB -- 3/30/06