Robert X. Cringely: Though you might not always know it from reading this column, PBS is a television network. And as a TV network, PBS is facing the same sort of technical challenges as its more commercial competitors. At this moment, that includes deciding how to play in the emerging world of digital downloads and IPTV.
But there is an aspect of this that most people don’t think about, and that’s the difference between national and local strategies, between how the network might want to run IPTV versus how local station managers see the opportunity. Up until now, IPTV has seemed to appeal more to the network than to its affiliates, but that’s just because people aren’t thinking clearly. IPTV might, in fact, lead to a renaissance in local television.
The Internet television story, even as written here in columns going back as far as the late 1990s, pushed the idea of enabling the aggregation of widely-dispersed viewing audiences, allowing programming to thrive that might not be successful on any local station, much less on the national network. A good example is NerdTV, which wouldn’t attract enough viewers on most PBS stations to even generate a rating, yet when offered as an Internet download, drawing from a global population, makes some pretty good numbers. But there is no concept called “local” in this aggregation model, so stations tend to feel threatened by it; if the network can reach local viewers directly, what need is there for a local station?
But it doesn’t have to be that way, because the supposed strengths of centralization aren’t really strengths at all when viewed in terms of the much more imposing issue of bandwidth costs, where all the advantages are local.
I explained this to a group of PBS station managers meeting last month in Orlando, Florida. Where these folks tended to fear IPTV portended the disintermediation of local television, I argued the exact opposite. My reasoning came down to the price differential between Internet bandwidth and intranet bandwidth, the latter being that bandwidth entirely within the ISPs local point of presence or data center. There is a lot more of this intranet bandwidth, for one thing. Depending on how their network is segmented, a local provider of cable Internet or DSL service may have gigabits of aggregate customer bandwidth attached to a much smaller Internet pipe. A 100-to-one ratio of internal to external bandwidth is typical, meaning the effective cost of internal bandwidth is 100 times lower.
What I advised the station general managers to do was to serve their traditional audiences as much as possible over internal, rather than external, connections. This means co-locating a server down at the telephone and cable TV companies, which isn’t hard to do since most communities have just two broadband providers, and the PBS station manager probably knows both of them from Rotary meetings or from the local United Way board.
In large part, this local advantage comes down to personal relationships. This is in stark contrast to my last two columns on Google -– a company that wants to avoid real human contact. Make friends with your local broadband providers, I said, then find a way to put your content INSIDE their network.
The advantages of this strategy are profound. Bandwidth costs go away completely, which not only frees up money for more programming or better servers, it becomes much more practical to display video with larger frame sizes, faster frame rates, and higher resolutions, creating a better viewing experience.
Imagine a local PBS station putting one of those lovely new eight-core Sun servers at the phone and cable companies. The servers would be packed with local programming and capable, through their multiple gig-Ethernet interfaces, of supporting as many simultaneous Internet viewers as the station serves broadcast users for many hours of the day.
What’s in this for the phone and cable companies is revenue sharing for advertising, as well as reducing demand on the ISP’s own Internet connection. They’ll understand instantly and see the revenue and cost-saving potential.
Put the station’s live signal on each server and enable downloads of past shows. Promote the idea of using the Internet for something viewers would like to see again. Offer video in 640-by480 at 30 frames-per-second. And do it all using unroutable IP addresses behind a NAT firewall with external interfaces, allowing only lower resolutions and smaller frames for true Internet viewers outside the ISP.
Of course, this has to be limited to content actually owned by the station, which generally means locally-produced content. But on PBS nearly all shows are, at some point, qualified as locally-produced. So do a deal with WGBH or WNET to offer their content (shows like Nova and Masterpiece Theater) on your servers. Absent a good peer-to-peer distribution scheme, this is the best way to get TV over the Internet and it is inherently local.
But wait, there’s more!
When John Warnock and Chuck Geschke started Adobe Systems almost 30 years ago, their goal was to build a printer — one printer — to serve the world. They knew the processing power required for a PostScript printer would be more than that of any computer they could buy circa 1979, so it logically followed that people wouldn’t own their own printers, they’d share one. It’s the same way the French look at their bread, which is never baked at home because the professional bakers do too good a job.
So John and Chuck set off to build what was essentially a service bureau for cool printing. This isn’t such an odd story, either, because I’ve heard similar tales from other pioneers who couldn’t imagine that their technologies would be eventually affordable by the masses so they planned centralized systems.
Generally, though, we tend to head the opposite direction and avoid centralization in computing. We live in the age of little iron, where large systems tend to mean clusters of smaller computers. But there are times when centralization is actually a good thing, where we don’t always have to be reinventing the wheel. One such area is video serving, which I think we’re handling in a remarkably stupid fashion.
If there is a 100-to-1 cost and quality advantage to intranet, rather than Internet, distribution, then we ought to do as much as possible without hitting the Internet, itself. This was the whole idea behind edge-caching and companies like Akamai. But I propose we take the model a step further and organize it in a different way: as a not-for-profit enterprise.
To the Libertarian powers of Silicon Valley, not-for-profit enterprises make little sense until they look at their own track records and realize that most high tech startups don’t show a profit for years, if ever. So what if you took away the profit incentive entirely, specifically for certain types of commodity activities like Internet video serving?
What brought this idea to mind was a contact recently fom a very well-known company looking for a cheaper way to distribute video. How did we do it for NerdTV? I explained the technique that keeps our costs at $0.16 per gigabyte, but petty theft just didn’t seem right for a really big operation. That’s when I realized that everyone is looking for a lower cost of video distribution while most ventures aren’t themselves, about video distribution. Some are, sure, like Akamai, but most aren’t.
What if we offered intranet video downloads and streaming in any format and resolution on a non-profit basis, which is to say something along the lines of cost plus three percent? It would be a huge hit! Every content business would hand over its basic bit schlepping because economies of scale would make the non-profit effort cheaper than they could possibly do themselves. The Internet video business would grow faster and fewer startups would die for technical or funding reasons. Local advertising in association with broadband ISPs would create new business models for content creators and ISPs alike.
And who would run this not-for-profit enterprise? In the U.S. it would be logical for it to be run by the 348 PBS local stations that live in every major and minor market, and already operate on a not-for-profit basis.
Instead of just Sesame Street, let them distribute the Sopranos and WWE Wrestling, too. And in doing so they can not only serve their local markets but those markets can help in new ways to support local programming.
The dream we have of a global network has kept us from realizing that when it comes to taking Internet video to the next level, our real heroes ought to be local.