Got called to a K-12 (educational) customer... they're having a high number of students and staff dropping from Zoom meetings, sometimes as many as 3 clients will drop in the same moment, but the rest of the room is unaffected. (They have ~15 kids in the room, each on zoom, with the remainder of the kids at home and rotate which kids are onsite each day.)
Doesn't happen from ethernet (and it's the same NAT/PAT, so not the firewall running out of translations or something). Affected users tend to be on 2.4ghz. Hotspots in the building are all served by APs on channel 1.
I'm aware that going to Wifi layer 1 for "Zoom disconnects"... feels like a stretch... but I found something weird: https://porktest.s3.amazonaws.com/etc/MysteryTrafficb.png
Most APs are seeing a 15mhz wide signal of what looks to me like FHSS traffic, centered on 2407mhz, below the channel 1 centerline. The longer I look at it, the more weird it looks. Its somehow less spread lower in the band and more spread higher. Its too narrowband to be Bluetooth (afaik).
There’s a separate ~2mhz wide carrier down near 2402mhz, unrelated. There’s another one at ~2430mhz. The solid lines I'm thinking are glass-break sensors.
- Its there at night, so its something resident in the building and not carried in by the students/staff.
- Most APs see it at roughly the same signal strength, meaning its widespread... but some APs barely see it at all. No real rhyme or reason.
- It wasn't there September 2019. The only significant change IT is aware of was UV lamps being added to air handlers, but that doesn't look like noise, that looks like FHSS traffic. (The glass-break lines and the 2mhz wide channels were there.)
What is this thing?
or
What obvious clue did I miss to point to the real cause of the Zoom disconnects? (It didnt occur on my test laptop while I was there, so no wireshark dump yet.)
No comments:
Post a Comment