So, just what do you do with a gigabit of bandwidth, anyway? Slate technology columnist Farhad Manjoo recently traveled to Kansas City, home of Google’s über-fast fiber optic Internet service — Google Fiber — to get a first hand look.
“During my time in Kansas City, I spoke to several local businesspeople, aspiring startup founders, and a few city boosters. They were all thrilled that Google had come to town, and the few who’d gotten access to the Google pipe said they really loved it. But I couldn’t find a single person who’d found a way to use Google Fiber to anywhere near its potential—or even a half or quarter of what it can do. It was even difficult to find people who could fully utilize Google Fiber in their imaginations.”
Okay, so much for the man-on-the-street interview angle. What about the point of view of the geeky engineers?
“This was true even of Google employees, both the folks on the ground in Kansas City and the execs who are managing Google Fiber from Mountain View, Calif. ‘What can you do with Google Fiber?’ I’d ask, and I’d often get an answer like, ‘Anything you want.’ Technically, this is true. It’s also singularly unhelpful.”
Let’s consider the possibility, then, that Manjoo is asking the wrong question. As fun as it is to instantly load and simultaneously play a half-dozen HD video streams, link up multiplexed teleconference sessions and upload every image in your high-school yearbook to Facebook, the point of gigabit Internet isn’t to saturate the bandwidth, any more than the point of buying a bigger home is to fill up its generous walk-in closets.
But bandwidth is only part of the story. Latency, too, figures in. Where your current broadband Internet connection may route you through a series of hops at rates of 50 to 300 milliseconds, Google Fiber might provide that same route at 1 to 3 milliseconds… provided that your target *also* has gigabit Internet. Which is to say, Metcalfe’s Law — which I’ve previously characterized as the second law that governs the Internet — has a role to play here:
“Metcalfe’s Law, offered by Bob Metcalfe — a guy who knows quite a lot about networks — he invented Ethernet, and founded 3Com… states that the utility of a network equals the square of the number of its users. Consider the computer you’re looking at now. Imagine it unplugged from the network. Alone. Isolated. It’s still a computer. You can run a spreadsheet, edit a document, play a game. But once you connect that computer to even just one more, the power of your own computer increases dramatically. You can now share those documents, or send messages to the other computer on your network. The utility of your computer continues to increase — geometrically — with each additional node that is introduced to your network.”
The corollary here: it’s not the sheer number of nodes on your network, but the number of nodes who have a similar capacity in terms of bandwidth and latency as you. Gigabit Internet will realize its full potential when *everything* you want to connect to is also served by gigabit Internet. Until then you’ll still be subject to the weakest link in the chain, or the longest (in terms of latency) hop in your route. Your connection to Netflix will be mediated by their server capacity, and the capacity of the routing networks in-between. Your peer-to-peer file or document sharing will depend on the rate and capacity of your peers’ networks. If ever there were an argument for cloud-based storage, where shared assets are kept in highly-connected and massively scalable data centers, this is probably it.
In the end, this isn’t entirely lost on Slate’s Manjoo:
“Gigabit broadband is like that. For it to become truly useful and necessary, we’ll need to see a long-term feedback loop of utility and acceptance. First, super-fast lines must allow us to do things that we can’t do with the pedestrian Internet. This will prompt more people to demand gigabit lines, which will in turn invite developers to create more apps that require high speed, and so on. What I discovered in Kansas City is that this cycle has not yet begun. Or, as Ars Technica put it recently, ‘The rest of the Internet is too slow for Google Fiber.’”
Let’s hope the rest of the Internet catches up. We’re waiting.