Exaflood still MIA according to latest Internet traffic data

The Internet is not about to explode (again). That's the most recent conclusion from the University of Minnesota MINTS project on Internet traffic rates. The site, headed by Professor Andrew Odlyzko, has been in the habit of debunking some of the more wild-eyed speculation about data growth—the Internet could drown beneath an exaflood of Hulu and P2P, ISPs need the freedom to discriminate wildly to keep the whole thing from blowing up Death Star-style, etc. Instead, growth remains fairly steady (and robust) at a manageable 50-60 percent per year.

New data points from major Internet peering points shows year-over-year traffic growth of anywhere from 24 percent to 65 percent, with the average in the 50-60 percent range. Such a growth rate has actually been seen for some time; the days of Internet traffic doubling every year (or even every hundred days) haven't been true for some time.

One outlier in the data was the PAIX set of Internet exchanges, operated by Switch and Data. PAIX points hosts many of the interconnects between US ISPs and those in Europe and Asia and are located on both coasts. Switch and Data notes in a recent announcement that its worldwide traffic growth estimate is 65 percent but that data at PAIX exchanges actually grew by 112 percent over the last twelve months.

This appears to be due almost solely to massive growth in the New York exchange, which grew by 295 percent this year. The Palo Alto exchange, which handles Asian interconnections, grew only by 67 percent.

Apart from the New York PAIX site, other numbers from TERENA, Cogent, Equinix, and various European peering points showed growth rates from 24 percent to 54 percent.

While significant, such rates won't destroy the Internet. Even at 100 percent year-over-year growth, Nemertes Research estimates that ISPs still have four or five years before demand exceeds supply—and that's only in access lines to customer premises. Internet backbone links still have no bandwidth problems and are not predicted to have any.

Assuming growth rates remain in the 50-60 percent region for the forseeable future, ISPs will have even longer to milk slower lines such as DSL. Cable operators have more bandwidth to use, and many are beginning to do so with technology like DOCSIS 3.0, which will give them even more time before they feel a serious crunch (uploads will remain an issue on cable networks, however).

But, under any growth scenario, it's fiber that remains the real winner. Fiber has unbelievable capacity that can be lit up easily by replacing the endpoint electronics and using more wavelengths to transmit data. As Nemertes notes, "fiber capacity is increasing more steeply than we projected last year and shows no signs of flattening. Fiber capacity is defined more by the electronics that illuminate the fiber than the fiber itself. Consequently, a sharply increasing capacity curve is reasonable since fiber interfaces will follow a Moore's Law dynamic with increasing capability per dollar spent."

Fiber to the home deployments, such as Verizon's FiOS remain in the best possible shape to handle future bandwidth challenges. And, as Verizon's example shows, "exaflood" horror stories should never be used in an attempt to justify a nonneutral 'Net on the grounds that the "tech simply can't handle it." The tech can handle it just fine; the question is whether companies are willing to handle the tech.