Generosity Is a Default Setting

International Day of Charity is celebrated on September 5, and the most generous aspect of many software programs is not a donation button. Instead, it is a default setting. Defaults determine what happens to users who never delve into the settings—something that applies to almost everyone. A default that silently consumes a resource a user lacks represents a minor act of unkindness multiplied on a large scale.

Conversely, a default that does not consume resources is a genuine gift that users do not have to request. One such change is worth considering. For many individuals, a phone is not merely a secondary connection; it is their primary means of accessing the internet. Tethering, the process of using a phone's internet connection to connect another device, like a laptop, is often slower than the phone's native radio connection.

This slowness is usually due to the laptop, not the carrier, and a simple solution is a default setting that requires no additional cost to modify. This feature detects the specific issue you are facing without consuming any of your data and the survey that informed its diagnostic method. The repository containing this tool can be found at https://github.com/xbill9/tether.

The tether-report utility reads all observable facts about the attached USB tether, applies a diagnostic algorithm, and generates a markdown report indicating the problems and the necessary changes. By default, it performs no measurements and incurs no cellular data costs. The standard measurement in this repository involves three single-stream transfers, a four-stream parallel test, and a ping—approximately 56 MB of metered data.

On an unlimited home plan, this is negligible. However, for prepaid plans purchased by the gigabyte, spending 56 MB to discover that the issue is a single-line setting is inefficient and a trade most speed-test tools force users to make. Thus, the tool separates the two processes. While the kernel already possesses all the necessary information—such as the driver, bus speed, negotiated link rate, MTU ceiling, NCM aggregation buffers, congestion control algorithm, interface error counters—it is free to access this data.

Most real problems are evident in this information. Users can opt-in to the more comprehensive measurement process, which will display the associated costs before executing the measurement. This diagnostic employs the same rubric as the survey, focusing on the following areas:

1. Single-stream: The link is operating at full capacity with a flat round-trip time (RTT).

2. Aggregate: The connection is limited by the USB 2.0 bus speed, which is approximately 480 Mbps.

3. Congestion control: The congestion control algorithm is not interfering with the radio connection.

4. Interface error counters: No errors are present.

The ideal scenario is when all four conditions are met, as this represents a free and effective solution. The tether-report utility spent no cellular data to achieve these four outcomes, demonstrating its efficiency. In the provided example, a Pixel 9a on Google Fi was tethered via USB-C, and the diagnostic process consumed no data.

The most critical issue identified was the presence of another route-capable link (wlo1) alongside the tether. Disconnecting this additional link resolved the issue. The tool cost zero cellular data while providing valuable insights into the tethered connection's performance.