The blog gives a hint at what might be coming. Included in the blog is a speed test from the home of a Google Fiber employee in Kansas City who is receiving 20.2 Gbps. I think this might be a signal to other ISPs that Google Fiber is prepared to surpass the capability of the XGS-PON technology that the industry is adopting. That technology delivers up to 10 gigabits symmetrical to a cluster of homes, depending on the electronics vendor. It’s obvious that Google Fiber is using something faster for the test than the currently available XGS-PON. It’s always been speculated that Google has developed its own customer electronics, but the company has always been moot on the issue.
It’s not easy for most current fiber providers to update to 20-gigabit speeds, even if there I a PON solution faster than 10 Gbps. An upgrade hits every portion of a network. It means a faster connection to neighborhoods. It means faster core routers and switches. And it means a more robust pipe to the Internet. Google Fiber has an edge in fiber backbone since it has built or leased dark fiber to many markets to support YouTube peering. It means all new ONTs and customer modems capable of receiving 20-gigabit speeds – and it means much faster WiFi within homes. Ultimately it means computers and devices cable of handling faster speeds.
Google Fiber was the first to make a national splash in 2010 with gigabit fiber for $70 per month – a price it has never increased. At that time, there were a handful of municipal, cooperatives, and small telcos that offered gigabit speeds – but all of them I know about charged significantly more than $70. It sounds like Google Fiber is getting ready to recalibrate the top of the speed market again. An affordable 20-gigabit product would certainly do that.
The most interesting thing said in the blog is that speed isn’t everything. The blog hints at having products that benefit from much faster speeds – something the industry has been searching for since the introduction of gigabit speeds. There are still very few uses that can fully utilize a gigabit connection in a home, let alone a much faster connection. There are some. I have a friend with several competitive games in the house that tax his gigabit FiOS connection. There are doctors with direct connections to hospitals that can use a gigabit to view complex imaging files. There are specialty engineers, data scientists, animators, and others who could use a gigabit and more if working from home. But most homes don’t use services that can use that much bandwidth.
The product on the near horizon that could use multiple gigabit bandwidth is 3D holograms as part of immersive virtual reality and telepresence. I keep waiting for somebody to offer such a product. It wouldn’t be hard to imagine hundreds of thousands of homes trying this almost immediately. My guess is that the roadblock to much faster services is the underlying middle-mile backbones. I don’t think most local ISPs have nearly enough backbone bandwidth to support multiple customers using a dedicated gigabit of bandwidth.
The other impediment to superfast broadband products is upload bandwidth. Telepresence is a 2-way service, and even if it can work on a gigabit download connection, there is no chance of such a service working on cable company networks where upload speeds are a minuscule fraction of download speeds. According to OpenVault, over 14% of homes now buy a gigabit download connection, but I have to imagine a large percentage of these connections are on cable companies.
It’s easy to write off fast broadband speeds as vanity purchases, and to some degree, it’s true. But the industry is now facing a classic chick-and-egg dilemma. We won’t get faster broadband products until there is some critical mass of homes ready to use them.
The blog says that Google will be discussing some of these issues in the coming weeks, including the upgrade of networks and maximizing speeds inside homes.