Which game is Google changing?

Richard Chirgwin analyses Google's decision to build its own gigabit networks, and comes to the conclusion that the chance to flex its muscles at telcos opposed to Net Neutrality could be one reason for the move.

Observers have suggested that Google's decision to build a network to connect end-users to the Net is "game changing." But what game is Google changeing?

The news part of the announcement is easy: Google says it’s going to build an experimental 1 Gbps network to “50,000 to 500,000” end users in America to examine “new ways to help make Internet access better and faster for everyone.”

Can we stop throwing our hats in the air and predicting doom for all other carriers that don’t follow Google’s example for just a minute, to look instead at what it means?

For example, let’s take on board criticisms of the NBN that we aren’t building a gigabit network here in Australia. That’s true – but the NBN is building a fibre network, and the fibres themselves don’t much care what speed you’re blinking the lasers at. That’s the job of the hardware, which can be upgraded.

NBN Co expects the equipment to be upgraded. Its CEO, Mike Quigley, said so to the Broadband and Beyond 2010 conference in Sydney, reminding delegates that in the 1980s we were still working with 2400 bits/s analogue modems.

But in the NBN – as in any commercial carrier – that upgrade would be on commercial terms, based on the customer’s willingness to buy an upgraded terminal, and on there being sufficient demand to install new switches and optical splitters.

It’s interesting to note that Google hasn’t put any price tag on its experiment. While it needs to be accountable to its shareholders for the spend, it doesn’t have the same constraints on it as the NBN.

By announcing an ‘experimental’ network funded by its other operations, Google gets to sidestep the commercial discipline that the network should be viable in its own right. That’s fine, as long as Google’s shareholders don’t object. But before the network is built, it’s premature to consider it a template for the whole world.

So what’s the upside for Google, apart from anything it might learn from the experiment?

In spite of the stated purpose of the experiment (looking at the applications that emerge over gigabit networks), Google will also get some financial lessons from the network, over and above the cost and effectiveness of new deployment technologies.

At the moment (in spite of the spin put out by the industry in America), Google’s cost of access to users is non-zero. It needs connections into the Internet, and while the outside world isn’t privy to how its connection costs are structured, those connections don’t come for free.

Google doesn’t get to build the network for free either: but once it’s built, Google doesn’t have to pay other companies to access those customers.

And it gets yet another lever to use in the ongoing argument in America about what constitutes “net neutrality”. If carriers start charging a “Google premium”, Google can hold out the threat that it will bypass the carriers entirely – at the same time creating a market for existing and new service providers to join the Google network (and help pay for any expansion that happens as customers of Google).

Would Google go beyond this experiment? It would depend, in the first instance, on how the financial question stacks up. We know from the global scale of Google’s private network that it is happy to pay money to bypass the telecoms industry where doing to justifies the cost. If the price of the access network is justified by a reduction in Google’s own access traffic, you can expect its access network to grow. If not, it will be treated as an interesting experiment – but not expanded.

For some people, this leads to questions about whether Google is becoming too big: would its influence in the search market, plus its “parallel Internet”, plus ownership of end-user access networks bring it under the ambit of American monopoly laws?

In a post-NBN Australia (and in any country with a significant fibre infrastructure) the question probably won’t arise.

Google’s not going to arrive in a future NBN-connected Toorak or Toowomba or Tempe and instantly devalue the NBN to zero, because the NBN’s owner/s (whether NBN Co under the government or under some future, post-spin-off structure) would be able to roll the gigabit upgrade far quicker than Google could roll a new fibre.

But it will provide plenty of excitement in the meantime.

Read more on WAN performance and optimisation