To fill up the entire Internet you’d need to flood the networks with 77 terabits per second of content, since that’s exactly how much capacity there is according to Telegeography. The analyst firm said today that while 77 Tbps of capacity is impressive, the total rate of growth in capacity is slowing to its lowest rate in five years thanks to the fewer new subscribers and the increasing use of content delivery networks.
We’ve covered the slowdown in broadband subscribers in the U.S., although there is an untapped market of roughly 19 million Americans who don’t have broadband — it appears most of them can’t afford it or just don’t want it. As for CDNs they are carrying ever more content and caching it at the edge, so fewer bits have to traverse the long haul networks. This saves the ISPs money and network investment costs.
Which means that instead of doubling (or more), Internet capacity around the world increased by 40 percent between 2011 and 2012. Happily for all of the ISPs complaining about not being able to meet demand, it appears that the growth of traffic is also slowing internationally at peak times. Telegeography says average international internet traffic grew 35 percent in 2012, down from 39 percent in 2011, and peak traffic grew 33 percent, which is far less than the 57 percent increase recorded in 2011.
But traffic, and thus demand for capacity, is still on the rise, which means no one thinks ISPs will have to stop investing in their networks, and long-haul cable companies should probably still be planning their upgrades. From the release:
Nevertheless, the underlying drivers of bandwidth demand remain strong. Broadband penetration rates in developing markets remain modest, leaving substantial room for new subscriber growth. In more mature markets, where the pace of broadband subscriber growth has slowed, faster broadband speeds and the growing adoption of bandwidth-intensive applications, most notably online video, are spurring higher traffic volumes per user.
Which does mean that ISPs will continue to invest and continue to complain about that investment. What’s not in this analysis is how much it costs to add new capacity to the network on a per bit basis. Telegeography has reported that long haul bandwidth costs are dropping, and part of that is because the equipment costs less, and it’s cheaper to add new capacity to the existing networks. It also costs different amounts to upgrade different parts of the network, making the price of adding another million bits an even more difficult number to find. Adding capacity at the edge may involve a new router or it may involve laying new fiber.
Regardless, those investments must continue to be made, but greater transparency on the cost of making them would benefit policymakers and even consumers trying to understand what their love of broadband costs.