Why IPTV Will Work

IPTV will work because the models embodied in YouTube and iTunes aren't the only models that will work. There is room for another model, which I'll get to in just a second, but first I want to provide some build up.

I did some quick research on Google about bandwidth costs and turned up some good work that somebody else has already done on the topic. I love that.

The above-linked blog goes some distance toward figuring out what the deal is with YouTube's money burn - the assessment concludes that YouTube is spending $1 million a month for around 6 petabytes of data transfer. That's 6 gigabytes for $1.

That's enough to serve quite a lot of videos.

It appears that traditional content provider networks are already trying to spread the fear, uncertainty and doubt about IPTV that always presages the death of a paradigm. Interesting, that.

I was doing a calculation (mind-blowingly complex) of how much it costs me, as a Dreamhost customer, to distribute data. I get a terabyte a month for something less than $10, but let's quote whole numbers to make it easy. That's

• $10 for 1TB


• $1 for 100GB


• $.10 for 10GB


• $.01 for 1GB


• $.001 for 100MB (that's 1/10th of a penny)


• $.0001 for 10MB (that's 1/100th of a penny)


• $.00001 for 1MB (that's 1/1000th of a penny)

You might be yawning right now, but this is really amazing to me.

Side note: one thing that is exposed here is a key component of every hosting companies offerings - that being an assumption that most customers won't come near to utilizing the full amount of bandwidth they are allowed. I don't think that Dreamhost could afford to sell a terabyte for $10 unless they understood that most customers won't use their full allotment. My question, and one that I'm sure won't be answered easily, is how much does Dreamhost oversell itself?

Overselling is commonplace in other Internet-related industries. Internet providers oversell access, assuming that their networks won't need to support the full number of subscribers at any given time. Cable companies in particular appear to be prime offenders at this practice.

So Dreamhost is extremely cheap, presumably because they oversell themselves. So far, I've seen no ill effects from this practice, if indeed this is their practice, so good for them. But what happens if I need all that bandwidth and more? Why wouldn't I just set up a second Dreamhost account for $7.95 and get another terabyte for less than $10? This instead of paying per-gigabyte overage fees of $1?

Back to the main story - why IPTV will work when so many say that the economics are fundamentally broken.

Traditional methods of Internet data delivery involve a server handling the needs of a client all by itself. Once the client gets the data they are free to move on. The idea of server mirrors simply makes several servers available, but does nothing to change this fundamentally inefficient approach. Both YouTube and iTunes use this approach.

Bittorrent has come along and introduced a means by which to harness media seekers as adjunct distributors of media. The rules of bittorrent require somebody who wants to get some media to also share some media. All files that are distributed get broken down into chunks so that during the download of any given file, a downloader will have many opportunities to share with others.

This approach not only works, but it scales AND it costs the distributor relatively little. In fact, the more people who use bittorrent, the better it works.

So, out there on the Internet at this very moment is a protocol that actually gets better the more people use it, at a pitance of what traditional media distribution costs. Sounds great, right?

The thing that bittorrent doesn't provide is individually crafted downloads per user - in other words, everybody gets the same file, and that's bad for the DRM folks. This appears to almost be a non-starter.

Except consider what happens when the model is designed for the broadest possible distribution of media. That's basically the model of traditional TV, so we have a frame of reference. Anyway, sponsorship of media that is distributed to an incredibly broad audience could become an enticing target to advertisers, just in the manner that advertising is attractive on traditional television, despite time-skipping technology (TiVO) etc.

So IPTV will be the same old TV we've always had, just delivered online to our computers using mass distribution protocols like bittorrent. What it will not be is a new way to get your NFL programming, as torrents don't do live programming. IPTV will coexist with traditional broadcast, perhaps some exclusive up-label content on iTunes, Netflix here and there, but it will go a long way toward ruining the hundreds of cable operations that exist today, reducing them to live news, talk and sports channels exclusively.