Policy by the Numbers: Telecom

White Spaces and Rural College Towns: the Beginning of a Beautiful Friendship

Monday, December 2, 2013

Michael Calabrese is Director of the Wireless Future Project at New America Foundation. Blair Levin is Executive Director of Gig.UFederal Communications Commission Chairman Tom Wheeler began his tenure by correctly noting, "Unlicensed spectrum has been and must continue to be the catalyst for innovation." With wireless traffic projected to increase 87 times in five years, we will increasingly depend on unlicensed spectrum to grow the mobile economy. Not only does unlicensed contain greater capacity—the capacity of Wi-Fi networks is 28 times greater than that of 3G and 4G networks—unlicensed enables huge savings. Wi-Fi off-load, projected to soon carry 60-80% of the traffic, will enable over 200 billion euros in annual savings for carriers in Europe alone.One promising area for unlicensed services is unused television channels, known as "White Spaces." White Spaces development, however, has been hampered by a chicken and egg problem. Without clarity about demand, manufacturers won't devote the resources necessary to lower equipment costs; without costs coming down, demand will not be sufficient.Several years ago we joined forces—Michael as an advocate for allocating White Spaces for unlicensed use, Blair from the direction of providing the high bandwidth networks rural college towns need and often don’t have—to create AIR.U, a project dedicated to using White Spaces to accelerate the deployment of next generation broadband networks in rural areas. AIR.U is a collaboration with the Declaration Networks Group and various Higher Education Groups, including the United Negro College Fund, the New England Board of Higher Education, the Corporation for Education Network Initiatives in California, the National Institute for Technology in Liberal Education, and Gig.U, as well as Google, Microsoft and the Appalachian Regional Commission.AIR.U network deploymentNow it's time for another step forward. Last month, Declaration Networks Group announced a "Quick Start" Network Program tailored for Higher Education communities to evaluate, design, and deploy high capacity broadband networks leveraging White Spaces. The program includes an assessment of expansion approaches and a sustainable path to increase the coverage and capacity of high-speed wireless connectivity to the community. The program is offered exclusively to the AIR.U institutions. It includes a network with a base configuration supporting multiple Wi-Fi hotspots, and a user group for AIR.U Quick Start participants to collaborate in developing White Space technology, establish best practices, and share approaches for community expansion activities. We believe this partnership between rural college communities and white spaces networks is the beginning of beautiful friendship, catalyzing more extensive deployments, and accelerating economic and educational development throughout the United States.

Measuring the spectrum across Europe

Thursday, February 14, 2013

Petri Mähönen is the head of the Institute for Networked Systems at RWTH Aachen University in Aachen, Germany. Janne Riihijärvi works as a senior research scientist in the institute.The debate over radio spectrum regulation has become one of the most intense and crucial telecommunications policy debates. Radio spectrum, in the sense of availability of radio frequencies for wireless data communications, is scarce. Wireless technologies have inherent limitations. Understanding the current radio spectrum use in different environments is important and harder than it might sound. First, as is now well understood, not all spectrum that is allocated is actually used. Second, spectrum measurements and realistic usage models are harder to make than people assume. Third, while regulatory databases and similar sources, such as white space databases, contain a lot of information how spectrum might be used, inevitable errors in propagation models significantly limit the value of this information. Finally, for unlicensed frequency bands and user-deployed networks, particularly emerging femtocells, there really is no substitute for in situ measurements.The policy debate needs to be quantitative and data driven, and several research groups around the world have been conducting different types of measurement campaigns on spectrum use for quite some time. Measurement campaigns by the Shared Spectrum Company are early examples, while longer-running efforts include the Spectrum Observatory of the Illinois Institute of Technology, as well as our own efforts in this domain. However, most existing campaigns have focused on measurements at a single location. Such measurements are invaluable to gauge the promises of new technologies, but ultimately lack statistical coverage. They also cannot support regulatory enforcement, national security applications, or dynamic radio environment based optimization.As part of the recently concluded European research project FARAMIR, we collaborated with industry and academic research groups to carry out extensive measurements on spectrum use across different urban and rural regions. During this work, we spent a week collecting measurements in different regions of London, UK, ranging from downtown shopping streets and tourist haunts to outer suburbs. Some 150 different locations were covered during the week, and at each location several hundreds of received power measurements were taken on different frequency bands. The first results were presented recently in the IEEE DySPAN 2012 conference in Seattle. Our measurements show that while the cellular bands are, as expected, practically fully utilized in urban environments, the usage of many other bands was significantly lower. Further, even the usage of cellular bands varied highly across the different measurement locations, falling to surprisingly low levels outside dense population centers. These results indicate that while availability of cellular spectrum in urban region is indeed limited, there seems to be little shortage of available frequencies in rural settings especially if low power transmitters are used. Overall the dynamics of spectrum use were observed to be very complicated as can be seen from the example below, illustrating spectrum use on the 900 MHz GSM downlink band at a location near Oxford Street in downtown London.

In addition to the results from our own analysis, we will gradually make the raw data from our measurements available to the research community. Last week we released the first data sets from the London measurement campaign, as well as our earlier comparison measurements of spectrum use across several different European cities. All of these data sets as well as documents describing results from our analysis can be obtained from the website of the FARAMIR project. In the future all the publicly available measurement data from the Institute for Networked systems will be made available through iNets Measurement Data Archive (currently under construction). We believe that sharing data from such measurement campaigns is key to developing global understanding on spectrum use, thereby providing solid foundation for making policy decisions for future spectrum use.

Modeling a Market for White Space

Thursday, November 29, 2012

Kate Harrison is a graduate student at UC Berkeley and Anant Sahai is an Associate Professor at UC Berkeley.Using TV white spaces means allowing wireless devices (e.g. wireless routers) to transmit on frequencies previously exclusive to over-the-air TV. The goal is not to eliminate over-the-air TV but instead to increase efficiency by maximizing existing resources. A useful analogy is pouring sand into a jar of large rocks, where the rocks in the jar naturally leave gaps for sand to fill in. We can think of the signals for TVs, called primaries, as the rocks, which leave room for signals from new devices, the secondaries, our sand.The principle concern is preventing harm to primaries. Secondaries must be "quiet" enough that TV sets can still "hear" TV signals (in communications lingo, the signal-to-noise ratio must not drop too much). Consequently, we must enforce a limit on the collective "volume" of secondaries.The standard approach is to hard-code the per-device limit on transmit power ("volume"). This works where devices have roughly the same requirements regardless of location. However, as the map below shows, white space availability varies greatly.

Large variation in the number of available white space channels.The natural response to a variable environment is to adapt to it. To be legal, white space devices must contact servers to register and get permission to transmit, which ensures they don’t get too close to protected TV signals (in this sense, the policy is already data-driven). With this setup, it’s easy to simultaneously assign a custom transmit power. We showed with data-driven simulations that there is a power limit function which allows significantly higher mobile data rates without hurting TV coverage:

Single transmit power everywhere

Transmit power varies by location

These maps were created in Matlab using US 2010 Census data by tract, the ITU propagation model, and a list of the 8,186 US TV towers, assuming white space ISP towers are placed to serve 2,000 people each. Find the code here.

Notice that data rates are much higher and more uniform in the variable-power map than in the single-power map. But an infinite number of functions satisfy the (linear) constraints of the problem (i.e. preserving TV reception). How should we pick in a principled way?The traditional economics approach is to assign prices (using real or pretend money) to transmit power and allow people to trade freely until everyone is satisfied. Given the quantity of wireless devices, this is practically infeasible: imagine asking people to manually adjust the power of their wireless routers or even determine their valuation for a unit of power.However, if we make the simple assumption that all devices (users) crave data rate, we can actually simulate their actions in a hypothetical market. This lets us approximate the optimal outcome easily without requiring any human interactions.In our award-winning paper, "Seeing the bigger picture: context-aware regulations," we created a proof-of-concept “market” under the additional assumption that fair access to white space services is important to society. For example, San Franciscans will need more per-channel power than Montanans because they have fewer available white space channels. This hypothetical market is just a min/max convex optimization problem which can be solved quickly using today’s data centers and scales well even with thousands of constraints.Since white space access already requires communication with a data center, we can easily apply changes there without deploying new white space devices. This lets us refine algorithms over time—including testing them in small regions before deploying them to the entire nation—in order to improve data rates for users. Through this, the white spaces could open up an exciting new realm of real-time data-driven policy.

The state of broadband Internet access in Kansas City

Friday, June 22, 2012

Google is all about building data-driven products. When users of Google Navigation wanted the ability to circumnavigate heavy traffic, we began to use real-time traffic data to suggest alternate routes. When Google engineers began to tackle the problem of automatic translation online, they relied on translation data from hundreds of thousands of websites in many languages.Similarly, as we're in the process of bringing Google Fiber to Kansas City, KS and Kansas City, MO, we need to first understand how many people here already have access to broadband, and more importantly, how many don't. In order to get a grasp on the situation we partnered with a group that's full of community experts—the Mayor's Bistate Innovation Team (MBIT)—to commission a study on broadband adoption and digital literacy in Kansas City. Today, we're gathering with MBIT at the Kansas City, MO Central Library to release and discuss the data that we've collected.The good news is that a lot Kansas Citians seem to recognize the value of the web. Those surveyed said that the Internet can be helpful when it comes to job hunting, getting health information, and learning new things.

Unfortunately, our study also illustrates that there is a real digital divide in both Kansas Cities. We found that 25% of Kansas Citians don't have broadband at home, 17% are not going online at all, and 8% are only using dial-up or slow speed wireless connections.These stats lead to a follow-up question: why are one-quarter of Kansas Citians not connected to the web at home? We found that one of the primary reasons is cost. 28% of those who don't use said that they don’t go online because they don't have a computer, or because Internet access is too expensive. Meanwhile, 41% of respondents said they don't go online because they just don't think it's relevant to their lives.This is a big deal. Using the Internet isn't just about checking email and social networking. Access to broadband—and knowing how to use it—has become essential when it comes to jobs, education, business and much more. The web provides a wealth of information and services for Internet users, and people who aren't online are, simply put, at a huge disadvantage. A job search, for example, is much harder today without the ability to review job listings and apply online. Similarly, computer skills and digital literacy have become perquisites for the majority of job opportunities.From a policy standpoint, we can try and address these issues by encouraging policies that will make computers and Internet access more affordable, and promote digital literacy initiatives.But a lot of outreach and education needs to take place on a community level. And that's why this morning we’ve joined representatives from amazing local nonprofits, schools, libraries, city governments and other community experts in a discussion about how to take action and get Kansas Citians online using broadband Internet access. We'll post the video of our discussion here as soon as it's ready. In the meantime, you can peruse or download the full results of our study.The Google Fiber project is about making the web better and faster. But it's also about making the Internet more accessible for people throughout Kansas City. Digital inclusion here is a priority for Google, and it's clear that it's also a priority for community nonprofits and the local governments.Posted by Kenneth Carter, Policy Counsel at Google

Understanding network performance, 460 terabytes at a time

Monday, March 5, 2012

In 2009, Google helped a group of researchers and industry partners launch Measurement Lab (M-Lab), an open platform for broadband measurement tools. Since then, the platform has produced 460 terabytes of open data to help users, researchers, companies, policymakers and others in the Internet community better understand the performance of broadband connections.

So what can one do with all that data? Because the data’s open, anyone can build on top of it, and that’s led to some exciting experiments.

We’ve been happy to see people using M-Lab not just to measure their own individual broadband connection, but also to make the aggregate data intelligible beyond the laboratory. In addition to a number of national regulators who use M-Lab tests and data to inform their decisions, individual developers and researchers are digging into the numbers and exposing the stories they tell. Below are just a few cool examples of what’s possible using 460+ terabytes of open measurement data in combination with some serious creativity:

During a Hack4Transparency event this summer, developers and designers working out of Italy and the UK created an interactive map in which users can correlate data from M-Lab’s NDT tests with other global metrics from Eurostat. In the screenshot below, we can see a correlation between M-Lab download speed and Eurostat data giving statistics on regular Internet use and population size. Not surprisingly, countries with better speeds show increased Internet use.

Tony Blank, a developer from Colorado, used data from M-Lab’s Glasnost tool, which helps users measure whether particular applications are being throttled by their broadband provider. His visualization, Glasviz, allows you to explore a dynamic, per-application view of the throttled applications worldwide.

Finally, Tiziana Refice, a researcher at Google, worked to join download throughput numbers from M-Lab’s NDT tool with OECD advertised speed data. What we see here is the top portion of a larger table, showing data for all countries (where it’s available). As the table demonstrates, in all cases the median measured download speed is well below the median advertised speed.

There are many, many more visualizations where those came from, and I encourage you to check them out and get to know some of the surprising and unsurprising conclusions that we’re able to make based on cool uses of good, open data.

posted by Meredith Whittaker, Program Manager for Google Research who focuses on network measurement

Broadband Competition: Same As It Ever Was

Thursday, February 16, 2012

Cross-posted from Andrew Crain's blog
Editor’s note: Andrew Crain is a telcom lawyer and scholar in residence at the Interdisciplinary Telecom Program at the University of Colorado. Because this blog explores examples of how data can drive decision making and improve the quality of our policies, this cross-post is especially fitting.
Yesterday, I talked about the 60 percent of Americans who have a choice between two broadband providers with the capacity to handle any services they might use. The picture isn’t as rosy for the other 40 percent. Most of them (30 percent of U.S. households) will also have access to broadband at speeds that can now handle any services currently available, but only from the cable company. The remaining 10 percent will not have access to these enhanced speeds from anyone.

Before we go further, we should recognize that this situation does not represent a sea change in the competitive landscape. While it is true that cable broadband is now several times faster than the fiber-to-the-node service from telecoms (100 to 25 mbps), cable modem service has always been 3 or 4 times faster than telecom broadband:

The real difference is that the telecom speeds are now sufficient to handle any services customers use. That wasn’t true in the past, when the speed differences had a real-world impact. In the ‘90s, cable customers could download music in a reasonable time and watch YouTube videos, while many DSL customers could not. A few years later, cable broadband customers could watch TV shows on Hulu, while many DSL customers could not. More recently, the same was true for HD videos and movies. Now, for 60 percent of Americans, both cable and broadband services are more than adequate for streaming HD movies.
And the fact that a large percentage of Americans can only get the top speeds from one provider isn’t new. In the early days of broadband, most households had access to service from only one provider (usually the cable provider). As speeds increased, many households had the choice of only cable service that could handle streaming video, while telecom broadband was well below 1 mbps.
The cable companies used this speed advantage in the marketplace, and in the past they had a larger share of broadband customers than they do now. The most recent statistics from the FCC are as of the end of 2010, when 56.3 percent (43.3 of 76.9 million) of broadband connections were via cable modems. That number has been increasing in the last few years, but cable’s percent of total broadband consumer connections still is not as high as it was in the 90s, when they had nearly two-thirds of the market.
Cable companies are again increasing market share, but it will be several years before they reach the levels they had in the ’90s. The FCC year-end 2010 report showed that the cable companies added 2.4 million broadband customers that year, while telecom companies added only 1 million. In other words, the cable companies gained 70% of the net broadband residential adds. The cable companies did even better last year. According to the Leichtman Research Group, the cable companies gained 67% of the new broadband wireline customers in the first quarter, 77% in the second quarter, and 83% in the third quarter. That sounds like a lot, but the focus on net adds is deceptive, because as broadband penetration has increased, the number of net adds has started to decline.
Why is this significant? Because the fact that net adds have decelerated means that if the cable companies gain a large percentage of net adds, their share of total broadband residential connections will grow slowly. Assuming that net adds level out at the 2010 level of 3.5 million per year, and the cable companies’ rate of gain increases to 90% of new customers, their share of residential customers will increase to 62.6% in five years.

That would be a significant gain, but the cable share of the wireline broadband market would still be lower than it was in the ‘90s. Even that number is likely overstated. It is more likely that net adds will continue to decline, which is only natural as the number of households with broadband increases to 70, 80 and 90 percent. Based upon current trends, the cable companies will probably add a couple of points to their percentage of total fixed broadband connections over the next half-decade.
It is possible that a fundamental change in the market is underway, and the 2011 net add numbers indicate that cable’s rate of gain is accelerating. But a fundamental change in market structure will not happen unless the telecom companies start loosing a significant number of customers. Whether that will happen is yet to be seen.
The fact that the competitive level of the broadband market hasn’t changed does not answer the ultimate question: is this level of competition good enough? And is is good enough for the 30 percent of Americans who have only one choice of very high speed broadband?
Posted by Andrew Crain

Everyone Already Pays their Fair Share

Wednesday, December 14, 2011

Six years ago, CEO Ed Whitacre declared that his company (then SBC, now AT&T) and cable operators had invested too heavily in the Internet’s infrastructure to let companies like Google, Yahoo! or Vonage “use the pipes for free.” His argument recently resurfaced in a report by A.T. Kearney, which claims that consumers don't pay enough for telephone, cable or other Internet service providers (ISPs) to continue building broadband infrastructure. The report also says that companies like Google should pay additional money to the ISPs in a “two-sided market.”

The truth is that no one is using the “pipes” for free. Most ISPs tell their investors that their contributions pay off and customers pay monthly fees for Internet access. Upon closer inspection of the data, we’ve drawn the following conclusions:

Consumers pay for Internet access because they want services like Google, YouTube, Netflix, and Yahoo! -- not because they are loyal to their ISP. 75 percent of consumers say they’re willing to pay for better service and higher data rates to improve the quality of streaming to their home.

Network operators are experiencing a downward pressure on prices -- a normal response in a market where unit costs are falling faster than consumption is growing. This demonstrates a successful competitive market rather than a failure, as A.T. Kearney claims. In addition, it validates researchers’ predictions that supply and demand for Internet access would grow at comparable rates and level prices -- thanks to Moore's law.

Companies like Skype that were once startups in a garage are now the driving force for consumers to purchase high-speed Internet. Broadband and mobile data providers made approximately €155 billion in Europe last year as a result of consumer demand for Internet-based content and applications.

A.T. Kearney’s proposal would result in European fixed network charges of approximately $1.5 billion for YouTube, which is, according to third parties, more than three times YouTube’s estimated global revenue of $450 million. Imagine the impact that this would have on new entrants (the next YouTube) which may not have the backing of a company like Google. It would effectively shut them out of the market.

To be clear, there’s something to be said for Whitacre’s sense of entitlement with fiber, copper and other miscellaneous attachments -- a sense of entitlement that I have felt, advocated for, and believed in. Prior to joining Google, I worked in telecommunications for almost 18 years, deploying infrastructure that makes the Internet work. The business is tough, expensive and heavily regulated because everything is installed on shared, public infrastructure. It takes a lot of people, machinery and coordination to get the networks built.

But there’s a line to be drawn. Regulators have done well to ensure that competitive providers have access to public resources like the Internet’s infrastructure, but we should be careful about asking them to dictate pricing structures for services that run over that infrastructure. Shouldn’t this be left to the consumers and the market?

posted by Patrick S. Ryan, Policy Counsel at Google