A step function improvement in capability is how Milo Medin described Google’s Kansas City fiber project at the February 13th IEEE ComSoc meeting in Santa Clara. That huge improvement in customer experience is in contrast to the incremental gains of MSO [Multiple System Operator] and telco broadband networks, which have much lower access speeds.
Medin, who is VP of Access for Google, described a Gigabit/second fiber network that eliminates the bottleneck between home and the cloud, unleashing new applications and devices both in the home and, by implication, throughout a city. Google’s incremental improvements in its construction and operations, its relatively simple offering and its grass-root marketing are as important to its success as its innovative fiber and home networking technologies.
The story of Google Fiber is pretty well-known by now; Google issued an RFI a couple of years ago to which 1,100 cities responded to be the test bed for Google’s fiber to the home project. What isn’t so well-known is that the motivation for this was the middling price/bandwidth performance of the U.S. as compared to other countries. Medin, who was a key figure in the early success of cable modems through his affiliation with @Home, suggested that, instead of complaining to government, Google decided to solve the problem. The unexpected response of so many communities was a surprise to Google and, according to Medin, an indicator of a pent-up demand.
Interestingly, government turns out to be part of the reason for their success, but not in the form of subsidies or tax breaks. The techniques Google and the local city are using to streamline the permit process and literally work together is saving an estimated 2% of the build cost. Similarly, attachment of fiber to the poles is made somewhat easier because the local utility is municipally owned.
Thanks a Bunch CEQA, No Google Fiber for California
Rules and regulations are definitely shaping where and how the service will develop. Echoing testimony before Congress, Medin suggested that as long as CEQA [California Environmental Quality Act] is in place in its current form, Google Fiber will be virtually non-existent in California (there is a 850 home Google FTTH project on the Stanford campus). The irony that Google’s home state will not see its fiber network anytime soon was not lost in the room full of engineers at the IEEE meeting.
Medin explained that anyone can use CEQA to initiate a lawsuit to block a development. He cited the example of the use of CEQA to delay the rollout of Uverse in San Francisco for years. A linchpin of Google’s approach is achieving scale at a fast-rate and the uncertainty caused by CEQA sinks their business case. And there is a business case, as Medin pointed out that the margins on broadband are as high as 95% for incumbent providers in urban areas.
To critics who suggest an infrastructure play is far afield for a “search” company, they should think again:
- With YouTube and their other Google properties, Google already operates one of the world’s largest Content Delivery Networks
- With a Fiber to the Home network, outside plant maintenance is almost zero, as compared to a traditional cable or telephone network.
- With a gigabit connection and customized hardware, the home becomes an extension of their data centers. Although it wasn’t said in his talk, they are sure to have TR-069 or equivalent technology to allow the monitoring of devices within the home. Additionally, network managed WiFi routers integrated into each set-top will deliver a better experience than the home WiFi networks cobbled together by consumers.
Google is taking an approach that, in some ways, is reminiscent of the old Ma Bell, whereby Google designs their own equipment. From Optical Network Terminals [ONTs] to set-top boxes, Google has created devices that maximize the customer experience [a DVR that records 8 programs at once] and minimizes operational cost. Medin indicated that Google has some of the world’s best optic engineers on staff.
Unlike the days of Ma Bell, Google can work with third-party manufacturers to build what they need, allowing them to introduce devices without the overhead burden of owning factories.
Keep It Simple Marketing
As with Google’s other offerings, they are taking a brand follows product approach to their fiber product. That is, the end service is the focus on creating an offer that provides great value and a high customer loyalty/buzz factor that will essentially market and sell itself. Like Google’s approach to their search home page, Google is keeping their offer simple. Unlike the Chinese food menu of a seemingly infinite number of tiers that traditional video and broadband operators offer, Google has only three tiers:
- 5 Mb/s with $300 construction charge (may be amortized at $25/month for 12 months) & no recurring charges for 7 years
- 1 Gb/s broadband with 1 Terabyte storage For $70 per month
- 1 Gb/s broadband with video for $120 per month with a Nexus 7 as a remote control
These three offerings probably cover 95% of the market. Amortized over 7 years, the 5 Mb/s tier exceeds the National Broadband Plan’s minimum at a very affordable rate of less than $4 per month and serves those who can least afford broadband. The $120 per month tier includes a basic level of video that many people would like. On a dollar per bit basis, the $70 provides great value to cord-cutters, while providing a superior broadband option for those who do not want to switch their existing video providers.
Further simplifying their offering is the decision they made not to offer telephone as part of their bundle. Although this decision was made for regulatory reasons, this reduces the operational complexity of their network and minimizes the staff required to run their network. With one less complex feature to offer, their network implementation is faster. They probably don’t lose much of their Total Addressable Market, given the number of people who are either wireless only or can easily pick a VoIP service (including Google Voice, which works great with a Obihai VoIP adapter).
Like what so many independent, rural operators have done with their Fiber to the Home deployments, Google is taking a grass-roots approach to marketing. Google uses a crowd-sourcing technique to determine where to build. Instead of taking a top-down approach that focus on demographics, Google split the Kansas City market into neighborhoods. When a critical mass of people commit to service in a given neighborhood, Google builds out that area creating what they call a “Fiberhood”.
Where they build is thus dependent upon the citizens of a given neighborhood. Like the way it has marketed its other Internet businesses, Google is betting on and seeding efforts to create a viral buzz about their network. One of the more interesting developments is their retail store. Although not mentioned in a recent Wall Street Journal article about Google’s rumored jump into retail, this point of presence offers a physical location to educate potential customers and the local influencers who will help sell their neighbors on the service.
And this approach seems to be working as Medin reported that in some neighborhoods 50% of the residents are committing to Google Fiber prior to build.
Just the Beginning
A gigabit to the home with its low latency and high-speed brings the compute power of the cloud to the home; particularly when much of the content is cached locally within Kansas City. In a sense, this extends Google’s cloud platform to the home and business, such that the performance at the end point is virtually the same as what it would be in the data center. Medin hinted that 1 Gb/s is just a start. It is not too difficult to imagine the types of things that could be enabled with this sort of bandwidth, such as:
- City-wide WiFi or some other wireless solution (Google has been received FCC authorization to experiment with various wireless approaches for access). City-wide wireless could offer a low-cost mobile/nomadic solution for its customers. It could also be important for autonomous transit options.
- Distributed data centers – with 1 Gb/s connections, a Peer to Peer compute network (think connection of those DVRs) becomes a possibility. Why not use the computing power as well and create a virtual data center spread over hundreds of thousands of residences.
- Like what Google has done with its Android and Chrome operating systems, the fiber network has the potential to enable applications from third-parties. It is possible that some of these apps might even come from existing telecom providers.
The Google fiber project in Kansas City is on its way to meeting its goal as a showcase of how low latency, Gigabit per second bandwidth can transform a city one neighborhood at a time. The fiber is really serving as a last mile nervous system that connects the seemingly disparate pieces to an ever-expanding Google ecosystem, which is where the change will really take place. Unfortunately for California residents, and particularly ironic for Silicon Valley residents, new Google Fiberhoods won’t be making their way to the Golden State anytime soon.
[Author’s Note: Thank you IEEE for the facilitating the excellent program that featured Medin as one of the speakers and thank you Alan Weissberger for your editing assistance].