Erratic logging behaviour

[rockwater]

Hi,
I'm quite new to WirelessMon, so please excuse me if I'm not doing things right. I'm using the evaluation version of WirelessMon V4.0, Build 1005. I experience the following problems:

When saving log files, I've noticed that not all the Access Points that are visible in the WirelessMon Summary Tab are saved in the APLogData file. Some SSID's are listed quite frequently (like the Access Point that I'm connected to), while others are completely ignored.

To make sure that at least the Access Points of greatest interest are included in the log file, I entered the names of these into the text box that's in the Configuration / Logging Options Tab. I've entered 7 names separated by semi-colons. This doesn't seem to help much, really. While all these 7 Access Points always stay marked as available in the WirelessMon Summary screen, only rarely are they all logged together within the same Sample No. But most of the time they aren't logged at all. I've also noticed that very often Sample No.'s are skipped while logging, so there are no records whatsoever entered at many the intended sampling moments.

This problem seems to happen quite chaotically. I can't find any underlying pattern. I've searched the WirelessMon Forum for an answer, but I seem to be the only person experiencing this problem. So perhaps I'm doing something wrong? I would be very glad to be able to iron this problem out.

rockwater

[Tim (PassMark)]

It might be the sample time is set too fast for your network card to keep up with the requests, especially if there are a lot of access points available in the same area, does changing it to only sample every 10 or 15 seconds help?

[rockwater]

Hi Tim,
Because in my neighbourhood there are lots of access points showing up in the WirelessMon Summary tab, I've set the sampling time to 60 seconds - only to prevent the log file from getting too bulky. My guess is that this interval should provide plenty of time for the log file to complete even with this many access points. It also doesn't seem to make much difference if I run WirelessMon minimized to the system tray, in Main Window or in mini Window mode. My computer is a tad slow, it has an AMD Athlon 2400+ cpu, and there's 1GB DDR ram installed on the mobo. I'm connected to the internet by a Linksys AE1000 wireless-n usb network adapter. But the hardware can't be the problem, I'd say.

[Tim (PassMark)]

Could you please run WirelessMon with the DEBUGMODE command line parameter for a few minutes (explained in more detail in the debug section of the help) and email a copy of the WirelessMonTestLog.txt file to us so we can check out what might be happening.

[rockwater]

O.K. I've mailed the WirelessMonTestLog.txt file as you requested.

[rockwater]

After a fresh reboot and restarting WirelessMon, logging seems to go much better now.
Sample numbers are no longer skipped and the list of SSID's that are indicated in the text box in the bottom of the Logging tab are all included in the APLogData file. So far so good.
If I leave the text box empty, however, still not all the SSID's visible in the Summary tab table are included in the log file. Clearing this text box requires WirelessMon to be restarted afresh, or else it will continue logging using these SSDI restrictions. That's how it happens with me, at least.

To give an example: there are 30 SSID's visible, 20 are available and 10 are not available. Sorting order is RSSI in ascending order. (First come the non-available SSID's, then the available SSID's in ascending order of RSSI strength.) Now 20 SSID's are included in the log file. This number is apparently equal to the number of the 20 available SSID's. But the log displays BOTH the available and non-available SSID's (for instance, 15 non-available and 5 available SSID's). The proportion of available vs. non-available SSID's varies seemingly randomly along the sample numbers, but the sample size per sample number always remains equal to the number of available SSID's. Why the sample size and the way it is subdivided are determined this way is not really clear, but it may be useful to know that WirelessMon behaves in this peculiar way.

If the listing on the Summary tab is sorted in descending order (SSID's with strongest RSSI signal first, followed by weaker signals, and finally the non-available SSID's), then all the available SSID's are indeed included in the log file and none of the non-available SSID's. So to get all the available SSID's logged, the list should be sorted in this way.

It took a little time to sort this out, but now the rationale of the logging module has become a little bit clearer. I would prefer, however, if the user would have more direct control over the way the log file is constructed. At the moment the user has control only over de sampling rate and over the list of SSID's that should be included in the APLogData log file. For instance, I would like to be able to make sure that the complete listing of all the SSID's in the main window of the Summary tab are included in every sample number of the log file. And NOT that a random selection is being made out of the collected SSID's. And there may be other rules that are used in the construction of the log files if a different sorting strategy is applied and that I haven't explored yet. It would be nice to have the log file display the exactly intended output.

[Tim (PassMark)]

Thanks for the extra information, I believe I've been able to replicate similar behavior with the sorting affecting the log so we should have a new build for you soon.

[rockwater]

I'm glad to have been of assistance. The sorting method that I used in the example is just one of several possible sorting orders. To give another example: When sorting on SSID name, WirelessMon acts likewise as when sorting on RSSI strength. Now again the sample size varies with the number of available SSID's. Because there is no inherent relationship between SSID name and availability, now the available and non-available SSID's will be randomly dispersed over the sample. WirelessMon simply starts selecting SSID's from the top of the list and eventually stops selecting when the sample size has become equal to the number of available SSID's. Not really a rational way to pick a sample. This way the log samples become a curious mixture of available and non-available networks. Networks that happen to have an SSID name low on the list have a much smaller chance of getting selected than if the name happens to be high on the list. And what's in a name? So I think it would be a good idea to have a thorough look at the way WirelessMon creates log files with all the possible sorting orders, to make WirelessMon create log files in a more rational way, so that users can benefit more from these log files. Perhaps it will even be better to give the user an additional option to have more control over the way the logs are created.

[Tim (PassMark)]

The new build is now available from http://www.passmark.com/products/wirelessmonitor.htm