Original URL: http://www.theregister.co.uk/2006/02/02/cuban_multitier_net/

Mark Cuban and the Chicken Little Netheads

Who will deliver me Video on Demand?

By Andrew Orlowski

Posted in Data Networking, 2nd February 2006 01:17 GMT

Net Futures Do Martians pay late fees? And why, in 2006, do Earthlings need to go to Blockbusters in the rain?

These are questions we need to bear in mind as background to one of the most contentious debates in US telecommunications policy today.

Our visiting, hypothetical Martian would be astonished to discover that with so much cable running under our streets, and popular entertainment already in digital formats, we must still trudge out and seek from a dowdy selection of physical product. Even Netflix requires a trip to the PO Box. Popular culture is brought to the home today inside a Macintosh raincoat, not a Macintosh computer.

The technical reason for this is because the internet is lousy at delivering video, of course. However, the latest attempt to alleviate this stupidity, and add a premium overlay for delivering high quality video quickly over IP, and so enabling the celestial movie jukebox, has caused great alarm. It's going to be introduced on networks such as AT&T Lightspeed, and Verizon's fibre Fios offering. But the prospect of VoD doesn't appeal to everyone.

"This is horrendous," is how former San Jose Mercury columnist Dan Gillmor described the idea recently.

"It's a threat not just to Citizens Media (sic) but to democracy itself."

What a pity that this important debate has degenerated into such hyperbole, and is now accompanied by irrational, sky-is-falling hysteria. There's also more than a hint of snobbery here, for internet people have never been comfortable with the idea of popular culture. Why, one may ask, should new forms of delivery of popular culture take a backseat to some strange technical abstraction - and a communications infrastructure that suits the internet bloggers - largely white, middle-class and media literate - at the expense of blue collar tastes.

It's a pity indeed, for there's a very rational case for concern here.

In recent years the Baby Bells, the incumbent owners of the physical infrastructure we use, have largely written regulatory policy to suit themselves. Just as the small family store is now a rarity, the death of the independent ISP is surely now only two or three years away. The "Bellheads", as Professor Rob Frieden of Penn State University characterizes them, are investing heavily in fiber capability which will solve a problem that the "Netheads" have proved themselves unable or unwilling to tackle, that of high quality video over IP.

So the Baby Bells have the pipes, are investing in improving them, and now want to monetize those investments. With an increasingly vociferous chorus led by AT&T, and former SBC boss Ed Whitacre, multi-tier pricing looks set to be introduced. At a premium, low latency guaranteed delivery service agreements will allow new applications - perhaps applications that save us a trip to Blockbuster in the rain - to be introduced.

Whitacre reiterated his case again this week:

"If someone wants to transmit a high quality service with no interruptions and ‘guaranteed this, guaranteed that’, they should be willing to pay for that,” he told the FT “Now they might pass it on to their customers who are looking at a movie, for example. But that ought to be a cost of doing business for them. They shouldn’t get on [the network] and expect a free ride.”

The dangers of AT&T abusing this, however, are immediately apparent. What's to stop a vertically integrated infrastructure provider deprioritizing every one else's traffic?

Mark Cuban, owner of the Dallas Mavericks basketball team, three High Definition TV networks, and the indie 60-cinema chain Landmark Theatres, acknowledges the concern as legitimate but says the Nethead lobby is holding back innovation and new services.

"There is a deserved undercurrent of fear that whatever the big networks, particularly the telcos do, will turn into a monopolistic and anti-consumer practice," he told us via an email interview.

But he says a network provider would be foolish to pull the trick on established internet content and service companies. It wouldn't work -

"If the telcos try to push Google to pay more, then they are idiots. Even Google Video has no need for tiered service. Nor does any other major content provider."

The whole emphasis on tiers of service is, says Cuban, misplaced.

"Tiers of service aren't necessary to enable streaming or downloads of movies. And they are probably the worst possible applications. Tiers of service are necessary so that applications that need priority can get it. The opponents have it backwards," he argues.

"Right now you have to pay a premium to get better service, and the telcos can pick and choose who to offer it to and at what price. But those aren't published standards. In fact the hows, wheres, and why aren't well documented at all."

"I would like to see tier of service, analogous to a network API, so that NEW applications that can benefit from reduced latency and guaranteed delivery, and possibly applications like multicast can be written."

These aren't just entertainment applications, he points out.

"I use the example of using an off the shelf high def camcorder to stream a realtime examination to a doctor at a hospital on a different segment on the network. That application shouldn't require the investment in private lines or a VPN, that's expensive and IT driven. I would rather see an open source version of the application that can be written, knowing that if a doctor is willing to run their end, and a patient is willing to run their end, the entire process can happen, at a service level that allows it to happen successfully, at a far lower cost than it would cost today."

"In the conversations I have with network providers, it's about enabling open apps that couldn't otherwise operate successfully on an open internet," Cuban told us.

Andrew Schmitt, Nyquist Capital, argues both the technical case and business case of a video-capable internet. But he also points out an important cultural difference between the "Bellhead" and "Nethead" engineers.

"The old AT&T culture - and I worked with them for some time - is that you can not lose a single bit. Ever."

"In the internet space, IP is designed to be a lossy protocol. Now they're faced with this explosion in video, and to throw away certain types of packets is a natural cultural response."

Under the Netheads' favorite abstraction - "End to End" - the network is dumb, and packets just find their way around.

"Under End to End it becomes virtually impossible to guarantee Quality of Service over multople networks," Schmitt argues, "but this can be easily fixed."

Like Cuban, Schmitt respects the case that the pipe owners can abuse their position - but he doesn't see it as rational, and he argues strongly against the regulation embodied in the "net neutrality bills".

"They built them, they own them, and if they shoot themselves in the foot - then they're only shooting themselves."

Cash rich companies such as Google could create alternative service offerings by building points of presence (POPs) at the edge of the AT&T network, he suggests.

So what's in the crystal ball?

History suggests two things. Firstly, that the Baby Bells, and the companies they invest in and operate, typically end up with the regulatory oversight that they're fairly comfortable with. And not always through some nefarious means, but simply because money buys power. Verizon reported an 2005 operating profit of $22bn on revenue of $75bn last week. As we noted, if you take the revenues of Apple, Google and Intel and combine them, you'd have something barely half the size.

Secondly, history also suggests that in the long term, very little gets in the way of new delivery systems for popular entertainment. If the nethead crowd want to pitch something as abstract as an "End to End" principle, or as semantically foggy as "Network Neutrality" against something that saves the public a trip to Blockbuster in the rain - it's going to be an uphill battle indeed.

So what should the future look like? We'd like you to tell us. ®