Volunteers Being Sought for Home Internet Study

CED Newsletter

The California Public Utilities Commission (CPUC), Geographical Information Center (GIC) at California State University, Chico and CSU, Monterey Bay are seeking 500 volunteers to take part in the CPUC Home Internet Study. The study aims to explore and analyze the Californian’s internet connections including:
* Performance of connections in rural areas, compared with connections in urban areas.
* Performance of DSL connections, compared with Cable connections.
* WiFi Router technologies in use in the California WiFi landscape.
* Performance of Ethernet (wired) connections vs. WiFi connections.

For additional information about the project, or to sign up, visit www.calspeed.net

I was a volunteer member of the beta test team for the CalSpeed data collection box. The box arrived in the mall, and I connected it to my internet router with a Gigahertz connection from Wave Broadband.

When Initially installed, before the beta test started,  the Wave 1Gig modem interface device’s speed ranged from 700 to 850 Mbps using my desktop Mac with Speedtest.com Speed checks were take at random times during the day. Good to go for the beta test.

Photo of beta test box


During the beta test, we were on vacation for two weeks in Seattle. While we were gone, PG&E change our electric meter. They cut off the power to the house while making the change. For some reason, the router quit working as did my drip irrigation system. I rebooted the router upon returning home and data collection restarted. After the reboot, I did a speed test on the Mac using SpeedTest.com and Wave’s Speed Test ranged between 300 Mbps and 500 Mbps.  Not the Wave promised 1000 Mbps!

These speeds are consistent with the overall averages collected by the CalSpeed data collection box. As you can see from the recorded data, my average was about 400 Mbps on ethernet and about 160 Mbps on WiFi.

Screenshot 2019-02-01 12.49.28
Screenshot of collected data from the beta test box on the CalSpeed webpage

The Wave network modem had built-in WiFi signals, one in the 2.4 GHz band and one in the 5.3 GHz band. It was not clear to me which band the CalSpeed box was monitoring, and I failed to ask.

I returned the beta test box to the development team, but I continue to monitor the Wave broadband signal with my DIY recorder box. Following the beta test, I downgraded my Wave connection to 250 Mbps Service, as my DIY box is limited to about 300 Mbps due to the circuit limitation on the Raspberry Pi processor board.

Raspberry Pi BB Recorder

Given all the marketing hype about broadband internet access speeds, the only valid method of determining the real speeds is field testing. Going out to the specific location and measuring the speed of the bits coming out of the ethernet port. So far, I am not getting what I am paying for, and there is a high probability that most users are not experiencing their ISPs advertised level of service.

Here is an example output:

Wave Tri Display

NTIA’s Request for Comments on Broadband Data Spurs Concerns About FCC’s Form 477

The Broadband Breakfast has the details. The following is excerpts:

According to US Telecom, the root of the problem is not “an inaccurate view of where broadband exists” but rather that “unserved locations are not mapped at all.” Because they are unmapped, the unserved areas remain excluded from policies to deploy broadband access.

“It is therefore our recommendation that NTIA concentrate its limited resources on augmenting the National Broadband Map with a more fulsome set of rural geocoded locations that may exist in the hands of other government entities,” US Telecom said.


The Wireless Internet Service Providers Association expressed concern that new efforts by the FCC to collect more granular data could press heavy demands on small service providers, who may have difficulty collecting such specific data.

“Data-collection efforts that seek broadband deployment at the sub-census block level will create new burdens on small broadband providers that will be ill-equipped to provide information at a more granular level,” WISPA said.

WISPA also noted that obtaining granular data can be costly and time-consuming for small business who may have to purchase new software or upgrade existing technology.

If you cannot recognize the problem, it is impossible to fix. Accurate broadband maps of unserved and underserved are essential to sound policymaking at the national, state, county and town level. A highly publicized crowdsourced program could identify target areas for detail ground truth surveys with recorders to documents the underserved conditions.  It is time to stop talking and start acting.

RCRC: Rural Broadband Update

On Thursday, Federal Communications Commission (FCC) Chairman Ajit Pai provided remarks at an event hosted by Axios.com. Chairman Pai covered a broad range of policy issues, including rural broadband and the digital divide.

Chairman Pai said his top priority for the remainder of his chairmanship is to close the digital divide, which he argues is the number one issue in internet policy today. The FCC is softening regulations and issuing funds to help small providers build out broadband networks in rural markets at a more affordable rate. On July 24, 2018, the FCC will administer the Connect America Fund Phase II (CAF-II) auction to provide funds that support broadband deployment in rural areas. CAF-II will provide $2 billion of funds to rural broadband programs over ten years to underserved communities.

Earlier this week, the House Energy and Commerce Subcommittee on Communications and Technology approved legislation that would establish a new office within the Department of Commerce to streamline federal funding applications for broadband deployment programs. H.R. 3994 would establish the Office of Internet Connectivity and Growth (OICG) to help rural communities get access to high-speed broadband. OICG will engage rural communities to develop best practices and strategies for promoting broadband access. In addition, the office will streamline the application process for broadband support programs across all federal agencies.

Source: RCRC Barbed Wire Newsletter [Emphsas Added]


Crowd Source Solutions – Validating broadband availability

In 2012 the Gold County Broadband Consortia, one of the 14 Consortial funded by the California Public Utilities Commission, recognized that the California Broadband Maps did not reveal the correct broadband coverage at advertised speeds. GCBC developed a survey and purchased booth space at the Nevada County Fair. They asked fairgoers to fill out the survey and report the availability of broadband at their address, including provider if they had broadband access and relative speed. About 400 surveys were filled out, with 334 that had useful information, including the phone number of the family filling out the survey. As the consultant on the project, I phoned many of the survey takers and validated the provided information.

The validated data was plotted using Arc/GIS Online and presented to the CPUC staff to demonstrate the deficiencies in the California Broadband Maps. The result was the formation of a CPUC GIS Working Group and creation of the CA Broadband Map Public Feedback Layer. The GCBC working with the CPUC developed a more comprehensive survey that encouraged the users to use online internet speed services to check their access speed and report it using the survey. The completed surveys were compiled on Excel spreadsheet and submitted to the CPUC. An online version of the form was developed that produced a Google Spreadsheet, which could be sent directly to the CPUC Mapping Team, after some minor clean up and validation. More details including plots of the findings HERE.

Seeking a more data-driven approach, I developed a broadband speed recorder based on an article in Make Magazine – Use Raspberry Pi to Measure Broadband Speeds to Hold Your ISP Accountable, February 1, 2016. [ https://makezine.com/projects/send-ticket-isp-when-your-internet-drops/ ] My goal was to build a recorder that could be deployed for less than $250 a unit. A box that worked when plugged into an ethernet port on a router and power applied, the only requirement was to check the ethernet status lights to ensure data was flowing to the box. After developing a prototype, I proposed to the GCBC they build ten of these recorders that could be loaned out for seven days and then returned for analysis. The data is downloaded from the Pi recorder in Excel format and analyzed using a Python data plotting program.

Recorder Sample Output

Wave Tri Display

Under the proposal, a week of data would be collected sampled every 30 minutes. This data would be summarized and plotted on a map using Arc/GIS Online, listing the min, max and average for ISP download, upload and ping timing in a popup information window. Under ideal conditions, 30-40 data points could be collected per month. The concept was to focus on areas with limited broadband validating the actual conditions of service, speed, and reliability, over a week period under varying conditions. More data points could be collected by building new Pi Recorders. The GCBC Program Manager decided not to accept the data collection proposal and continued to collect online surveys.

The Federal Government had a similar program called Measuring Broadband America.

The Measuring Broadband America (MBA) program is an ongoing nationwide performance study of broadband service in the United States that developed out of a recommendation by the National Broadband Plan to improve the availability of information for consumers about their broadband service. This program is built on principles of openness and transparency. The FCC has made available to stakeholders and the general public the open source software used on both its fixed and mobile applications, the data collected, and detailed information regarding the FCC’s technical methodology for analyzing the collected data.

Measuring Fixed Wire Broadband

Reports offer results of rigorous broadband performance testing for 13 of the largest wireline broadband providers that serve well over 80 percent of the U.S. consumer market. Tests conducted used automated, direct measurements of service delivered to the homes of thousands of volunteers across the United States. The Measuring Fixed Broadband studies began in 2011 with the release of annual reports based on data typically collected during a single month with few large-scale traffic events, such as major holidays, sports events or other elections. The data analyzed in the Reports thus reflect stable network conditions that provide the most accurate view of a provider’s performance under controlled conditions.

The FCC works in collaboration with SamKnows, an international statistics, and analytics firm supporting similar projects in other countries around the world. The Measuring Broadband America program incorporates the latest engineering best practices from these diverse stakeholders to collect and report the most accurate data for consumer broadband performance in the United States. Learn More. See most recent report HERE.

The latest report is from December of 2016, and we are almost to the middle of 2018. The last data on file is August of 2016. It appears the data collection has been terminated. Final report charts are HERE.

Other than the testing being terminated, the real problem is that it was only testing wireline services (telephone wire, fiber, and cable) when fixed wireless ISPs serve significant portions of rural areas. Wireline services are in dense population areas and are not meeting the needs of more dispersed rural citizens. If these rural citizens have broadband at all, it is crappy DSL or fixed wireless services. DSL was tested, but fixed wireless was not. Mobile wireless (cell phone) service was tested, and coverage reported. Mobile testing covered in a future post.

If there is a nationwide test program initiated to improve broadband maps, to identify where investments should be focused, it should be on the fringes where the fixed wireless ISPs are struggling to fill in the gaps left by the top tier wireline providers. The broadband testing needs to focus on collecting and plotting the scope of fixed wireless services using low-cost recorders such as those provided by SamKnows or DIY reordered like the Raspberry Pi device described above.

If I were a community leader in a rural community seeking broadband access, I would be crowdsourcing the data collection, not waiting for a government testing program. Those communities who have done their homework and have identified the broadband deficiencies will be prepared to respond to State or Federal request for proposals quickly. Expanding and improving fixed wireless will have the more significant return on investment for the American Taxpayer and rural citizens than giving more money to wireline providers who are reluctant to serve low-density population communities or rural areas.

Sen Manchin, Makes Wise Mapping Decision

The FCC says Sen. Joe Manchin (D-W.Va.) can challenge its decision to exclude some areas of his state from subsidies to expand mobile broadband. The Mobility Fund auction will award $4.53 billion to companies to offer service in rural areas, but the FCC said some communities already have adequate service and are thus ineligible for the funding. Manchin, though, says he wants to do his testing to prove the agency got it wrong about coverage in his state. “We conclude that there is good cause for granting the requested waiver,” the FCC said in an order Friday, citing Manchin’s “long record of engagement” on wireless issues in West Virginia.

Source: POLITICO Morning Tech

Smart move by the Senator. “Manchin, though, says he wants to do his testing to prove the agency got it wrong about coverage in his state.”

If the FCC is relying on the Broadband Maps to make their coverage decision, there is a high probability they will be wrong every time. The maps are not accurate; they are based on claims made by the providers and not ground truth reality. Early in the development of the CA Broadband maps, the Gold Country Broadband Consortia demonstrated the coverage maps were not connected to the real world. More details HERE and HERE.

The FCC was equally wise to grant the waiver request; they also know how inaccurate the broadband maps are, and what Senator Manchin’s test would reveal.

The real solution is in the field testing. Here is my answer to field testing, up to 100 Mbps.

Raspberry Pi BB Recorder
DIY Raspberry Pi low-cost broadband recorder 


Wave Tri Display
Graphic analysis os recorder box file

DIY or Go Without Broadband: Part Three

Part One  and  Part Two

There are multiple online tools for testing internet speeds, in addition to the CPUC apps mentioned in Part Two. Some example here:

These tools test internet broadband or WiFi networks for download speeds, upload speeds, connects and pings. The download value tells you how much data a WiFi or broadband connection can move from a server to your computer. The upload value indicates how quickly you can transfer files from your computer to a server in the cloud. Upload is essential for those who regularly use cloud services to store data. Connects shows you the number of connections that can be established simultaneously. The higher the contact number, the better. Pings refer to the response delay, which is mainly attractive to gamers, lower values are preferable. The maximum possible data transmission speed depends on whether the user has broadband, fiber, WiFi or another connection, as well as the service provider pricing plan.

You can run these test any time of the day or night to check the speed of your network. It has been my experience; the systems start slowing down after school gets out in the afternoon and right after the dinner hours as people start download or streaming movies.

The major providers also list speed tests on their web pages.  But, are they accurate or do they fudge the results.  I have no idea, and that is why I wanted a dedicated recording device.  A device which can automatically test the speeds at regular intervals over long periods of time.

It eliminates the need to remember to do the testing during high use times. And, it provides a data record that can be discussed with the provider when a problem occurs, with precise times when the problem shows up.   

I started to build my recorder the day Make Magazine arrived in the mail, Use Raspberry Pi to Measure Broadband Speeds to Hold Your ISP Accountable, February 1, 2016.  I was already experimenting with Raspberry Pi computers, having built a device to locate RF noise which as creating radio interference in the house.  I had on hand all the components to make a recorder, except for the software which I could download from the sources listed in the article.

The author Alasdair Allan indicated the time required to build the device was 1–3 Hours and lists the difficulty as easy.  However, the article assumes that the builder is familiar with the Raspberry Pi and command line programming, making this a geek project.  The goal of Alasdair’s project was to hold an internet service provider accountable for advertised service levels. 

My goal was a modification of author’s, I wanted to build a recorder that could be plugged into the internet with an ethernet cable and power supply and start recording, creating a week of 24/7 data which could be downloaded for analysis.  A plug and play recording device that the Gold Country Broadband Consortia could loan to citizens so they can record their broadband speed and reliability.  The initial plan was to build ten of the devices and loans them out for a week at a time. The data analysis by GCBC, the information consolidated, plotted and then forwarded to the CPUC.

Raspberry Pi BB Recorder
Recorder box, Ethernet cable, and power wart

The GCBC proposed speed test monitoring to the CPUC in an extension to their work plan, but the proposal was not specific about the process and equipment to be used. Once the grant awarded to the Sierra Business Council, which had taken over the GCBC program management from SEDCorp, the program manager showed little interest in using the recording device which was in beta testing.  The ten loaner recorders were never built.

A sample of the recorder output when loaded into an Excel Spreadsheet.

Recorder Sample Output

At present, it is one of a kind recorder which I loan to friends and family to test their internet speeds.  According to contacts at the CPUC they are building a similar recording box capable of recording up to Gigabit speeds. The latest version of the Raspberry Pi  B+ has a Gigabit Ethernet port, but the processor speed limits data collection to about 300 Mbps. This is well beyond the minimum criteria set by the FCC and CPUC and could be useful in validating neighborhood internet speeds.  I am planning to build a next-generation recorder with the latest Raspberry Pi board to test the speed limits.  Stay Tuned. 

DIY Resources:

Make Magazine Article: Use Raspberry Pi to Measure Broadband Speeds to Hold Your ISP Accountable

Raspberry Pi Kit: Amazon CanaKit Raspberry Pi 3 B+ (B Plus) with Premium Clear Case and 2.5A Power Supply

GitHub, speedtest-cli-extras:  https://github.com/HenrikBengtsson/speedtest-cli-extras

Plug and Play Box:  Send me an email russ@russell-steele.com if you are interested in a plug and play box.