Man, InSSIDer can cause some real problems

[azazel1024]

Its a great testing tool for surveying, but it has some serious issues with it.

To my dismay I have found

1) With my tablet if I have InSSIDer open, network throughput plumets from around 9-10MB/sec to 2-3MB/sec. Close InSSIDer and it jumps back up. If I have my SSIDs for 2.4 and 5GHz set to the same one, InSSIDer on my tablet also likes misreporting the 5GHz signal strength. It jumps around by ~20dBm not moving anything around. Seperate SSIDs, no issues.

2) With my laptop if InSSIDer is up, no speed issues on transfers. HOWEVER, it REFUSES to roam between access points and also has other strange behavior, like if InSSIDer is open and I disable/reenable wifi, my laptop will always want to associate with the access point that is running 5GHz, even if the signal strength is -90dBm or so and I am right next to one of the other access points with ~-30dBm 2.4GHz signal strength (only if I set it to prefer 5GHz band). InSSIDer closed and the laptop roams seamlessly and effectively immediately between access points (within about 2 maybe 3 seconds max once it gets near enough to a better signal strength AP. My tablet takes maybe 5-7s and my phones/wife's iPad 2 take more like 8-12s).

Just...bizzare behavior. I took me an hour of trying to figure out what was wrong with my laptop and why crazy crap was happening with refusing to roam and stuff.

Good news was it forced me to play with drivers a lot while I was trying to figure it out. The Intel 7260ac has some serious issues with TP-Link routers/APs Rx-ing from them (Tx is fine) with the newer drivers. Original Windows 8.1 16.0.36 (MS signed) drivers works great. SUPER fast. In playing I found that the latest drivers which will work really well speed wise are the 16.5.3 drivers (Intel's initial 8.1 release drivers, which take a little bit of finding on the internet as they aren't on Intel's site any more).

 

[thiggins]

Reviewer Scott DeLeeuw found the throughput impact while testing wireless extenders. Had him going for awhile, too!

 

[L&LD]

I initially used inssider 'religiously' when I first discovered it. However, I quickly learned that what was indicated as 'best' was not always the best channel to be using for many of my customers.

Now? I just do my throughput tests just like I used to do anyway with inssider before.


As we know, the act of testing may change the outcome of what we're testing in the first place...

With inssider, this is at a whole different level.

 

[stevech]

I make a habit of never putting test & monitoring software on the same computer/device that's functioning in a performance test. This always colors the results by some degree.

E.g., a Netgear GS108 ethernet switch which can mirror ethernet frames between two ports to a third port where one has a PC running Wireshark as a passive observer.
There is software that does the same concept, passive "sniffer" software for WiFi that can run on a dedicated device (promiscuous mode for 802.11 frames, irrespective of protocol). One I've owned/used was called AirMagnet - a startup, then acquired by Cisco, then sold to ?

 

[azazel1024]

I make a habit of never putting test & monitoring software on the same computer/device that's functioning in a performance test. This always colors the results by some degree.

E.g., a Netgear GS108 ethernet switch which can mirror ethernet frames between two ports to a third port where one has a PC running Wireshark as a passive observer.
There is software that does the same concept, passive "sniffer" software for WiFi that can run on a dedicated device (promiscuous mode for 802.11 frames, irrespective of protocol). One I've owned/used was called AirMagnet - a startup, then acquired by Cisco, then sold to ?

Excellent suggestion.

For testing, yeah, I always do throughput testing. That is my only standard for siting things. I've mostly been using InSSIDer to test things like which AP and and which band my device is connected to as I been doing a lot of roaming testing recently and also a bit of tweaking AP locations for better connection and roaming experience (especially as I am new to 5GHz). It didn't really occur to me that it could be causing issues with roaming.

 

[sfx2000]

Inssider works within the Windows Driver frameworks, and it forces the Wireless NIC to tune away much more often to get the RSSI values and to look for additional AP's.

So yes, there is a performance hit when running InSSIDer on the primary network interface - if you have more than a single interface, InSSIDer will allow you to use one for scanning while the other is unaffected.

sfx

 

[stevech]

AirMagnet ... now owned by Fluke, an excellent company.
They have a free download eval version - but only certain WiFi chipset/cards and OSes all all of its features to work.

Its function that I used way back on many projects is measuring channel utilization - by SSID, percent of channel time/capacity used. I'd let it run for hours or days at a prospective locale of APs. Evenings, weekends. That then gives a statistically valid assessment.
http://www.flukenetworks.com/content/wifi-analyzer-pro-product-demo-video
Channel utilization is shown in that video - top left graphic.

Of course, you can do such an assessment then someone comes along ant puts up an always-on streaming WiFi security camera and changes things. But such a camera isn't as bad as streaming 1080p at home (or trying to do so!).

AirMagnet may be expensive to own.
They also sell AirMagnet Spectrum - which is a spectrum analyzer using a special hardware receiver USB dongle. It's pricey.

 

[remixedcat]

I found that it has affected speeds as well on several adapters. I normally just open it, get the dBm value, close it, then run the test software.

I haven't tested Xirrus's tool while doing the tests and its UI is a little wierd looking. Netsurveyor is another good scanning tool but it doesn't display the manufacturer of the AP it does display the BSSID so you have to manually look for the OUI. Which is a pain.