Robert Cringely: I spoke recently with an old friend who is a bandwidth broker. He buys and sells bandwidth on fiber-optic networks around the world. And he told me something that I found not completely surprising, but I certainly hadn’t known: Google controls more network fiber than any other organization.
Inside a Google data center
This is not to say that Google OWNS all that fiber, just that they control it through agreements with network operators. I find two very interesting aspects to this story: 1) that Google has acquired — or even needs to acquire — so much bandwidth, and; 2) that they don’t own it, since probably the cheapest way to pick up that volume of fiber would be to simply buy out any number of backbone providers like Level 3 Communications.
Google loves secrecy. That they’ve been acquiring fiber assets hasn’t been a secret, but the sheer volume of these acquisitions HAS been. Why? One thought is that it kept down the price since people didn’t really know it was Google snatching up this stuff (they’ve done it under a number of different corporate names). But if price was the issue, then why hasn’t Google just bought the companies that own the fiber? It made no sense until I scratched my head and thought a bit further, at which point it became obvious that Google wants to — in its own way — control the Internet. In fact, they probably control it already and we just haven’t noticed.
There are two aspects to this control issue, but let’s take the legal one first. If Google bought a bunch of Internet backbone providers, such a move would of course get the attention of regulators from the U.S. Department of Justice and the U.S. Federal Trade Commission, the two federal agencies charged with looking at large corporate mergers for signs of anti-competitive activity. But simply acquiring legal control of those same assets through leases and other long-term contracts doesn’t trigger such an examination, though perhaps it should. By renting instead of buying, Google was able to acquire its fiber assets primarily in secret. The game was over before most of us even knew there WAS a game.
The second aspect of this is the whole idea that the game is already over for control of the Internet. I touched on this concept back in 1998 when I wrote my first column about PayPal, which at the time had been offering its core service for less than a year and already had eight million members. I wrote then that PayPal had already won the Internet payments race, which time has since showed they had. PayPal’s confidence was based on analysis of its own growth. Understanding the potential range of growth, looking at the rate of subscriber acceleration, and using second derivative analysis of these data, PayPal was pretty darned sure, even back in 1998, that its competitors at the time would never be able to catch up.
Topix.net founder Rich Skrenta recently took a similar approach to argue that Google, like PayPal, has already won the game and represents to most users the face of the Internet. Skrenta (in this week’s links) argues that Google’s dominance of search and advertising is so profound that most competitors — especially Yahoo — would probably be better off NOT even attempting to compete and simply let Google handle search and advertising while Yahoo provides content. He’s probably correct. Skrenta argues that even if services come along that are superior to Google’s, in order to become dominant they’ll have to overcome Google’s brand recognition with users, which is almost impossible to do. So just being better than Google isn’t enough.
All this is prelude for understanding what Google intends to actually DO with all this technology, which I have only lately begun to figure out.
I live in South Carolina, a state that I can argue qualifies as a technology backwater despite being the shrimp and grits capital of the world. Why, then, are the local business pages filled with stories about Google preparing to build massive data centers here? Google is apparently negotiating to build data centers in Goose Creek, a town not far from Charleston, where I live, in Columbia, the state capital, and a third location across the border in Georgia. To read the papers, Google might choose one or another of these locations, but according to people I have spoken with who are fairly close to the action, Google actually seems intent on building in all three locations.
Why would Google need two data centers in a state with only four million residents? Why would they need to buy 520 acres in a Goose Creek industrial park when that’s probably 100 times as much land as any conceivable data center would require?
Google is building a LOT of data centers. The company appears to be as attracted to cheap and reliable electric power as it is to population proximity. In Goose Creek they bought those 520 acres from the local state-owned electric utility, which probably answers the land question posed above. By buying out all the remaining building sites in an industrial park owned by an electric utility, Google guarantees itself a vast and uninterruptible supply of power, much as it has done in Oregon by building a data center next to a hydroelectric dam or back here again in Columbia by building near a nuclear power station.
Of course this doesn’t answer the question why Google needs so much capacity in the first place, but I have a theory on that. I think Google is building for a future they see but most of the rest of us don’t. I’ll go further and guess that Google is planning to build similar data centers in many states and that the two centers they are apparently preparing to build here in South Carolina are probably intended mainly to SERVE South Carolina. That’s perhaps 100,000 servers for four million potential users or 40 users per server. What computing service could possibly require such resources?
The answer is pretty simple. Google intends to take over most of the functions of existing fixed networks in our lives, notably telephone and cable television.
The Internet as we know it is a shell game, with ISPs building their profits primarily on how many users they can have practically share the same Internet connection. Based on the idea that most users aren’t on the net at the same time and even when they are online they are mainly between keystrokes and doing little or nothing when viewed on a per-millisecond basis, ISPs typically leverage the Internet bandwidth they have purchased by a factor of at least 20X and sometimes as much as 100X, which means that DSL line or cable modem that you think is delivering multi-megabits per second is really only guaranteeing you as much bandwidth as you could get with most dial-up accounts.
This bandwidth leveraging hasn’t been a problem to date, but it is about to become a huge problem as we all embrace Internet video. When we are all grabbing one to two hours of high-quality video per day off the net, there is no way the current network infrastructure will support that level of use. At that point we can accept that the Internet can’t do what we are asking it to do OR we can find a way to make the Internet do what we are asking it to do. Enter Google and its many, many regional data centers to fill this gap.
Looking at this problem from another angle, right now somewhat more than half of all Internet bandwidth is being used for BitTorrent traffic, which is mainly video. Yet if you surveyed your neighbors you’d find that few of them are BitTorrent users. Less than 5 percent of all Internet users are presently consuming more than 50 percent of all bandwidth. Broadband ISPs hate these super users and would like to find ways to isolate or otherwise reject them. It’s BitTorrent — not Yahoo or Google — that has been the target of the anti-net neutrality trash talk from telcos and cable companies. But the fact is that rather than being an anomaly, these are simply early adopters and we’ll all soon follow in their footsteps. And when that happens, there won’t be enough bandwidth to support what we want to do from any centralized perspective. A single data center, no matter how large, won’t be enough. Google is just the first large player to recognize this fact as their building program proves.
It is becoming very obvious what will happen over the next two to three years. More and more of us will be downloading movies and television shows over the net and with that our usage patterns will change. Instead of using 1-3 gigabytes per month, as most broadband Internet users have in recent years, we’ll go to 1-3 gigabytes per DAY — a 30X increase that will place a huge backbone burden on ISPs. Those ISPs will be faced with the option of increasing their backbone connections by 30X, which would kill all profits, OR they could accept a peering arrangement with the local Google data center.
Seeing Google as their only alternative to bankruptcy, the ISPs will all sign on, and in doing so will transfer most of their subscriber value to Google, which will act as a huge proxy server for the Internet. We won’t know if we’re accessing the Internet or Google and for all practical purposes it won’t matter. Google will become our phone company, our cable company, our stereo system and our digital video recorder. Soon we won’t be able to live without Google, which will have marginalized the ISPs and assumed most of the market capitalization of all the service providers it has undermined — about $1 trillion in all — which places today’s $500 Google share price about eight times too low.
It’s a grand plan, but can Google pull it off? Yes they can.
Google’s two computing centers, each the size of a football field, in The Dalles, Ore.
Via I, Cringely