This is the resolution (I hope!) of:
That hopelessly long thread describes weeks of frustrating attempts to understand why the Wi-Fi throughput of my V1 Microsoft Surface Book, maybe a year ago, began dropping to zero every ten seconds. It happened on any access point in any mode with any configuration, and if the Surface Book was awake the speed of other machines connecting to the same access point would also be reduced on the 10 sec cycle.
Wireshark showed the Surface Book setting the IEEE 802.11 "PWR MGT: STA will go to sleep" flag every ten seconds, missing four packets, ignoring the AP's retransmissions and RTS packets, and eventually after dozens of DUP ACKs and retries getting re-synchronized. But no clue why.
Microsoft Message Analyzer, which is an amazing tool for studying problems with any aspect of Windows, provided clues. Learning to navigate its overwhelming selection of options and quirks is not easy. It doesn't duplicate the information you can see in Wireshark at all, you need to run both simultaneously and match timestamps to understand how they relate. But here's what happened every ten seconds:
Despite knowing it was useless to do an AP scan - "No auto switch for the current connection (TP-LINK_2.4GHz_CA8583)" - it would scan every ten seconds anyway!
22274 None 2019-04-16T20:17:25.2290175 RpcCall Scan from client 3780
22275 None 2019-04-16T20:17:25.2290500 Scan for networks. Interface = Marvell AVASTAR Wireless-AC Network Controller, scan type = 0x80000003, flush BSS list = True
22276 None 2019-04-16T20:17:25.2291499 Begin Scan
--> Note the ~5 seconds wasted out of every 10! <--
26245 None 2019-04-16T20:17:30.0942333 Number of Unique Wlan Networks 3
26246 None 2019-04-16T20:17:30.0943993 No auto switch for the current connection (TP-LINK_2.4GHz_CA8583).
26247 None 2019-04-16T20:17:30.0945666 Scan completion Status 0
26248 None 2019-04-16T20:17:30.0945725 Number of Unique Wlan Networks 3
26259 None 2019-04-16T20:17:30.1009674 Number of Unique Wlan Networks 3
26265 None 2019-04-16T20:17:30.1034665 Number of Unique Wlan Networks 3
26266 None 2019-04-16T20:17:30.1034931 Number of Unique Wlan Networks 3
26267 None 2019-04-16T20:17:30.1040412 Number of Unique Wlan Networks 3
26276 None 2019-04-16T20:17:30.1102738 Number of Unique Wlan Networks 3
And it did that very clumsily, literally killing the connection and cutting the overall throughput to less than half of optimal.
On May 2 it suddenly stopped doing that. It now displays this message once per minute:
224165 None 2019-05-03T11:48:53.3685359 Microsoft_Windows_WLAN_AutoConfig Discard this round of background scan because the current connection does not allow auto switch. Interface = Marvell AVASTAR Wireless-AC Network Controller
And it now scans the active channel only when first connecting on wake from hibernation. Wireshark shows other stations doing frequent probes of the same channel, with no lost packets or slowdowns, but the Surface Book is no longer doing any repeating probes.
So what changed? I don't know! The "netsh wlan show all" command shows only the same "AutoSwitch: Do not switch to other networks" it showed before. I "Agent Ransacked" the entire Surface Book for files changed between the two different situations, and found nothing conclusive. There are four Windows "Remediation" logs spanning the change, but the one that includes network related events, though it shows management events back to April 30, only shows network detail after the wake on May 2 with the problem gone. The others show nothing I can recognize... But Microsoft is masterful at hiding things behind cryptic ID values...
I do vaguely remember reading that Microsoft had released a remediation update to clean up obsolete and contradictory network settings, but they show no updates installed here since April 10 and 18. The other Remediation.etl files in C:\Program Files\rempl\Logs seem to follow earlier known updates, so I'm suspecting Microsoft silently released yet another example of "Windows as a Service" over which we're allowed no control.
I was hoping to understand the problem, but I'll settle for having it magically gone!
Finally ! hurrah !
It sure would have been nice to know what caused the problem. But hey - it's working now. Every so often that's all you can hope for.
I suggest you make a back up of the entire system as it is right now ,and save it forever.
Also, very carefully document every configuration chnage you make in the future, regardless of how insignificant it seems at the time.
I'm betting that there were some conflicted settings, somewhere, that even though they should be unrelated, actually interfered with each other.
I haven't had to use the message analyzer in a long time (we had better diagnostics in our client devices), but I guess I should go back and refresh my knowledge there.
Thanks very much for discussing your problem on the forum.
Definitely time for a serious backup! And yes, I'm logging more and more of my actions in a searchable calendar, along with keeping about 170K lines of web history in NirSoft BrowsingHistoryView. Which combines history from all browsers, and amazingly logs the local files I open as well as web activity.
For those new to the Microsoft tools, Microsoft Message Analyzer seems a big step up from the older Microsoft Network Monitor. Here are a few good clues I found:
--> Or use New Session -> (New Data Source) -> Live Trace -> Select Scenario
--> Do not scroll or click! Type search term, click chosen scenario
I barely scratched the surface of what can be done there.
Thank you Howard for your company on this adventure!
Glad to be of help.