FCC to Test 'Absolute' Internet Speeds

slfisher 1 499 Views

In response to criticism of the way it was testing broadband Internet speeds, the Federal Communications Commission, following up a promise it made last month, it had hired a vendor to help it perform more accurate tests.

"In a couple of weeks, we will be asking for consumers from across the country to voluntarily install hardware in their homes (on an opt-in basis) that is capable of measuring broadband performance," said the FCC's Dave Vorhaus, an expert advisor in economic opportunity, on the FCC's blog. "The measurements will give us results across a broad swath of providers, service tiers and geographic areas."

The FCC is working with a UK company called SamKnows that has recently performed similar testing there, Vorhaus said. "The FCC will also release a Public Notice in the coming days with details on SamKnows’ technical approach and methodology to allow for comment and new ideas," he added. The company will receive about $600,000 for the work, according to an article in the Wall Street Journal.

Other companies that bid on the contract, which will create five to 10 U.S. jobs, included the Nielson Co., the Wall Street Journal said. SamKnows UK said it would be licensing its technology without charge to a new American company that will be set-up, and based in Washington, specifically for this project.

Approximately 10,000 people are expected to volunteer for the speed test. "The specially developed ‘White Box’ does not monitor the home network or web traffic, but solely focuses on the relevant ISP’s network." according to the SamKnows . "In addition people who volunteer for the project are given access to their own data so that they can track the performance of their broadband connection comparing it against what their ISP is billing them for."

That should be interesting.

Dani AI

Generated

As observed, a move to volunteer-installed, hardware-based measurements changes the quality of broadband data compared with one-off browser tests. Hardware boxes and dedicated probes reduce variability from browsers and background apps, but meaningful results require consistent deployment, clear metrics, and sane interpretation. The notes below offer practical checks and repeatable steps that make volunteer data useful for long-term comparisons and for troubleshooting an individual connection.

Essential metrics and what they reveal: sustained throughput (not just a short burst), goodput (application payload after protocol overhead), one-way latency, jitter (latency variation), and packet loss. Short browser-based tests often overstate peak speed (TCP burst effects) or understate real-world performance (wireless losses, CPU limits). For real-time apps, jitter and loss matter more than raw Mbps; for bulk transfers, sustained goodput is the key number.

Quick checklist and reproducible tests to run before accepting any single measurement as “truth”:

  • Use a wired Ethernet link for the primary measurement whenever possible and isolate the test endpoint from other traffic.
  • Repeat tests at multiple times of day (peak vs off-peak) and record timestamps.
  • Capture basic network traces and control-plane info (link rates, modem channel/bonding status, NIC stats).
  • Controlled test commands (examples):
# start a server elsewhere:
iperf3 -s

# from the test host (download test)
iperf3 -c SERVER_IP -P 4 -t 30 -R

# quick loss/latency checks
ping -c 50 8.8.8.8
traceroute 8.8.8.8

Interpreting problems: if measured performance is persistently below the plan, first verify the local layer (cables, splitters, modem/router logs, firmware), re-run tests wired with the modem in bridge mode if possible, and test to multiple remote servers to rule out peering/CDN effects. Preserve logs and timestamps when contacting the ISP. Finally, review the measurement project's privacy and data-retention policy and watch for sampling bias: volunteer panels can oversample tech-savvy areas or certain equipment, so aggregated conclusions need documented methodology to be defensible.

Be a part of the DaniWeb community

We're a friendly, industry-focused community of developers, IT pros, digital marketers, and technology enthusiasts meeting, networking, learning, and sharing knowledge.

Edit Preview