A little over a year ago, my colleagues and I modeled current and projected Internet capacity and anticipated demand to see if the curves ever crossed. The short answer: Yes, sometime before 2012, access (though not core) capacity would no longer be enough to serve demand.
This was the first time anyone had ever looked at demand and capacity independently, and also the first time that anyone had applied a "Moore's Law of Innovation" to the process of demand generation. Essentially, we made the assumption that human creativity is unpredictable and boundless -- as long as the 'Net continues to exist, people will dream up new ways to use it. So, just as Moore observed that the density of transistors on a chip will continue to increase exponentially -- without knowing or caring about the precise manufacturing techniques that enabled each step of the increase -- we noted that people will come up with networked applications, without predicting specifically what the next application will be.
At the time, our research was greeted with no small degree of skepticism, but in the past 12 months we've seen clear validation. In particular, moves by carriers and cable companies to throttle back heavy users have indicated that these companies are feeling the heat from demand that exceeds their access capabilities.
Since last year, for example, Comcast has implemented a monthly usage cap of 250GB. Though 250GB sounds like a lot of traffic, I've spoken with Comcast users who have exceeded the cap twice in the past six months and came within 15% of the cap in three of the other four months, thanks to downloading video from the Olympics, the Democratic and Republican national conventions, and daily online system backups. (And no, this usage does not include porn).
Though this traffic load is more than typical, it certainly isn't exceptional. Keep in mind that biological systems, such as user adoption of new technologies, tend to follow Gaussian curves. So today's exceptional users are tomorrow's mainstream users -- meaning that this type of usage will become typical over the next three to five years. The fact that Comcast's network is, by the company's own admission, not able to cope with such usage patterns is a clear indication that the crunch we predicted last year is beginning to occur. As Cisco put it, "Today's bandwidth hog is tomorrow's average user."
We recently updated the model with new data. No surprise: Evidence increases that demand will exceed access capacity within roughly the next four years, although capacity in the core (including raw fiber capacity and switching/routing capability) continues to outpace demand.
But that's only half the story. This year, in addition to looking at raw bandwidth capacity, we took a look at what I'm calling "logical issues": challenges to Internet scalability such as addressing, routing, multihoming and mobility. The news there is even worse: In a nutshell, Internet scalability is rapidly reaching limits due to some architectural issues inherent in the design of the 'Net. To see how, stay tuned.