ISS crew Wi-Fi investigation

The ISS crew uses tablets on the station's wireless network, and the tablets kept disconnecting. I was given a list of device hardware addresses, several days of access point logs and one question: what is happening to these connections?

I expected the internship to be mostly coding. Most of it was reading logs.

At first I pulled events by hand for each tablet: first appearance, connection attempts, authentication, successes, failures and disconnects with their reason codes. When more devices were added, my co-intern and I wrote a Python script to extract those events.

I plotted them as a color-coded timeline with one lane per device, colored by access point and band, so the team could see the patterns directly.

• A tablet stuck reconnecting on a fixed interval for more than half an hour
• Bursts where many devices disassociated at once
• Connections that completed every step except the last. A privacy feature had given the tablet a randomized hardware address that the server didn't recognize, and the feature turned itself back on whenever Wi-Fi was toggled
• Sessions that ended when screens turned off

Over three days we ran a test series on two tablets: screen timeouts, logins with and without the privacy setting, wrong passwords, walking out of range and loading one access point with many devices.

Several theories were ruled out. Typing speed made no difference. One tablet's configuration profile explained its screen-off disconnects. We also found that one consumer phone could make an access point reboot after a few minutes. Each result went into a report and a timeline for the engineers.

Power supply threshold testing on an access point, Ethernet cable repair, and one afternoon in the Mission Evaluation Room listening to the crew install an Ethernet cable live.