Bandwidth Schmandwidth

It's been one of those "good news, bad news" months for bandwidth. At the Cisco Analyst meeting Chambers and other all mentioned an IP Traffic Forecast that Cisco had assembled (and has made available on their Web site along with a companion piece). For those of us who have been frustrated trying to get any meaningful traffic data in the past this is a wonderful document that presents a coherent view of what's happening to IP traffic in the Internet (hint: it's growing!). On the other hand, Chambers got exuberant and suggested that in fact video usage could cause traffic to grow much faster citing Cisco's own internal network growth driven by their aggressive use of TelePresence and other video services. Always eager to steal keen insights on the future I tried to find the facts under the enthusiasm and came up short. It's a longer story than fits here but this is the Readers' Digest version. (1) Video On Demand is clearly a happening thing. (2) Unfortunately the evidence that exists suggests that what people will do with VoD is time shift the same stuff (it's not one of them mythical "long tail" things). In which case, (3) the video needs can be served with content caches of some form right near the point of consumption and all that bandwidth won't do much to drive network equipment spending. If any of you dear readers has a more compelling argument why video use will drive a lot of core Internet bandwidth please share it with me!