The Need for Speed
Progress Snapshot
Release 4.10 April 2008
by Bret Swanson*
View as PDF
High definition (HD) content is hitting the Net even faster than I thought it would, with ongoing implications for Internet policy. In a series of articles[1],[2] and reports[3] over the last 18 months, I predicted an "exaflood" of Internet video and rich content, including the arrival of HD Internet video. In January 2008, co-author George Gilder and I estimated that between 2006 and 2015, U.S. Internet and IP traffic would grow 50-fold. HD video, we predicted, would be a large component of this surge and would require some $100 billion new U.S. network investment over the next five years.
With enough ingenuity and risk capital, we can surmount the technical and financial challenges of the exaflood. But not if Washington places additional regulatory obstacles in our path. Today, the Senate Commerce Committee is once again exploring the possibility of regulating the Internet in ways that could discourage or even prohibit the rational pricing and technical mechanisms needed to deliver high-end services. These possible legislative or agency actions thus potentially threaten the successful deployment of the new networks required to deliver applications like HD video.
Although tantalizing as a new consumer application, we believed that Net HD was probably still a few years away. After all, fuzzy-screen YouTube is still in its youth, broadband is still in the deployment phase, and who could forget our 20-plus year wait for HD television? New technology companies, however, have already proved us too conservative. Some Net-based video is approaching HD quality even now. We should remember our own credo, not to underestimate the power of the digital broadband Net compared to the old, analog, broadcast, but narrowband world.
For an example of the new HD offerings, one might look to Move Networks. With $46 million in new venture money from Benchmark Capital, Comcast, and Cisco, Move delivers high-end content via variable-rate "streamlets," from several megabits per second when available, down to several hundred kilobits per second when either your connection or some other network node is feeling lazy. Move's easily downloadable media player offers a nice demo[4], and Move's customers – ABC, Fox, ESPN, Discovery Channel, etc. – are delivering DVD-quality and near-HD-quality content right now. Meanwhile, PluggedIn Media has just begun streaming a large selection of some 10,000 near-DVD quality music videos licensed from Universal Music Group, EMI Group Ltd., and Sony BMG Music Entertainment. The Move-enabled picture on my computer appears crisper than many of the HD pictures on my widescreen HD television, depending on the channel and program.
But there is a key factor: I am using a dual-core computer with 4 gigabytes of memory and a new Nvidia graphics processor. I am also, crucially, using a fast broadband connection, which allows Move to deliver a constant 2.3 megabit stream. In narrowband settings, Move reverts to a lower 300 kilobit stream made familiar by YouTube. Broadband is thus still the key to making high quality video work.
Last week The Wall Street Journal profiled the competing broadband strategies of the big communications providers. On the one hand were Verizon, with its fast-but-expensive fiber-to-the-home vision, and Comcast with its Docsis 3.0 upgrade plan. On the other hand were AT&T and Time Warner, with their less fleet, but more frugal network plans.
"When it comes to Internet access," the Journal asked, "is there such a thing as too fast? That's a question U.S. Internet providers are grappling with as they place strategic bets on whether or not to upgrade their networks to offer high-priced, superhigh-speed Web connections."[5]
Verizon says equipment now in the field lets it boost broadband speeds to more than 100 megabits a second. What's more, last year it started adding new equipment that will allow for speeds of as much as 400 megabits a second, says Verizon spokesman Eric Rabe. "We think having too much bandwidth is a little like being too rich or too thin," says Mr. Rabe. "For most of us, it is difficult to be too much of either."
Comcast CEO Brian Roberts is equally optimistic.
"Broadband was instrumental to the success of Google, Amazon, eBay, YouTube and all other graphics- and video-rich Web services we now take for granted ... So when we boost Web speeds 10, 20, maybe up to 50 times faster than what you're used to today, it will mean a whole new world of innovation that we can barely imagine."
But are Verizon and Comcast expending lots of money and energy for nothing? Is this a case of the tortoise and the hare?
AT&T spokesman Michael Coe says consumers may not even be able to benefit from the very high speeds since congestion in other parts of the Internet network can ultimately slow downloads.
John Dvorak of PC Magazine argued a similar point.
In today's world, bandwidth demand is similar to what processing demand was 20 years ago. You just can't get enough speed, no matter how hard you try. Even when you have enough speed on your own end, some other bottleneck is killing you.
This comes to mind as, over the past few months, I've noticed how many YouTube videos essentially come to a grinding halt halfway through playback and display that little spinning timer....
All too often, it's not the speed of my connection that's at issue – it's the speed of the connection at the other end. It may not even be the connection speed itself; it may simply be the site's ability to deliver content at full speed under heavy demand.[6]
Whether delays are caused by narrowband last-mile links, poky web servers in the data center, or too many latency-inducing TCP/IP hops in between, the conclusion is that we need more bandwidth and thus more investment across the entire network, from Googleplex data center in the core to last-mile fiber at the edge.
We also need more sophisticated parsing and prioritizing at many nodes of the Net. Last week NetworkWorld asked, "Can your network handle HD video?", and highlighted HD's many technical challenges.[7]
Sven Rasmussen, a LAN/WAN specialist at CDW in Vernon Hills, Ill., says a good first step is to increase LAN speeds to 1G with 10G uplinks to core infrastructure....
Rasmussen recommends installing content filters for applications that aren't mission critical, as well as for certain types of data and access to unapproved Web sites. Next, companies should roll out traffic-shaping and QoS tools that let them compress, cache and prioritize traffic. "Video and voice are two types of traffic that if they're not received in perfect order, won't make any sense, so make sure those applications get the most bandwidth they need vs. e-mail and HTTP, which can be retransmitted," he says. IT teams won't have to worry about separate tools for traffic shaping and QoS much longer, he says. "With high definition and IPTV gaining traction in the next five years, it'll be commonplace that all networking equipment will feature [QoS] and traffic optimization tools," he says.
Policy-based networking, content-delivery networks, traffic management, 10 Gigabit Ethernet upgrades, massively scalable content servers: this is what consumes network engineers as they try to deliver fast and robust service to their company's employees or their customers. Silicon Valley and the global telecom and silicon industries are consumed with the very same challenges and are right now deploying the new bandwidth, traffic management, and high-throughput server technologies that will make HD video and interactive services a more compelling and satisfying experience.
Today, HD video channels are challenging capacity on many cable TV networks.[8] Some consumers have noted that the picture quality of some HD channels is not as good as others. They believe extra compression is being used to stuff the channels into bulging coaxial networks. Three, instead of the usual two, digital HD channels are being compressed into one old analog channel, they say.
New architectures like switched video and new compression protocols like MPEG 4, not to mention more raw bandwidth, will help service providers deliver the full HD experience more of the time. Abundant fiber-optic and wireless bandwidth will transcend most of the petty fights currently driving the policy debate.
But inevitably, as new content and new channels proliferate, and as we someday migrate upward from mere HD to UltraHD or whatever the newer variant may be, certain links and nodes on the network will be crowded or even overwhelmed. Choices will have to be made. Which bits go first? Which channels make the cut? If we allow, these questions will be easily resolved by (1) sophisticated network technologies that ensure robust customer service and (2) pricing plans that match the interests of consumers and service providers.
But if Washington insinuates itself into this complex, dynamic, growing, organic Net, all bets are off. We could, once again, end up waiting two decades for the next generation services that should be just around the corner.
