Rating Wi-Fi Stability

[dksadiq]

Hello all,

Firstly, thanks to the guys behind SNB for the services & assistance provided via the site & these forums.

Secondly, to my question:

Does anyone know of any criteria (standardized or otherwise) for rating Wi-Fi stability? I've checked the certification programs offered by the Wi-Fi Alliance but was surprised to find that it seems there are no tests on basic Wi-Fi properties like throughput and stability.

I've also come across a few cool tools like NetStress and iperf/jperf to measure and plot the throughput of a Wi-Fi connection measured over a period of time. From the results, one could obtain the average and standard deviation of the throughput and I was hoping I could find some criteria that would rate the connection as Very Stable, Stable, etc based on these two values (or some others) - but it seems such criteria is non-existent!

I'd really appreciate if someone could help me out with this or in case it unrealistic to develop such a rating scale, can someone please explain why?

Thanks a lot.

Sadiq.

 

[thiggins]

There is no such thing right now and probably will never be. There are simply too many variables to provide a meaningful measure.

That said, IxChariot generates a 95% Confidence interval number that could be part of what you are looking for. It's been on my to-do list for some time to add that information to our Charts database.

But that would be for a single client in a perfect environment (no other wireless traffic). The number probably wouldn't be useful for predicting real-world performance.

 

[stevech]

Hello all,

Firstly, thanks to the guys behind SNB for the services & assistance provided via the site & these forums.

Secondly, to my question:

Does anyone know of any criteria (standardized or otherwise) for rating Wi-Fi stability? I've checked the certification programs offered by the Wi-Fi Alliance but was surprised to find that it seems there are no tests on basic Wi-Fi properties like throughput and stability.

I've also come across a few cool tools like NetStress and iperf/jperf to measure and plot the throughput of a Wi-Fi connection measured over a period of time. From the results, one could obtain the average and standard deviation of the throughput and I was hoping I could find some criteria that would rate the connection as Very Stable, Stable, etc based on these two values (or some others) - but it seems such criteria is non-existent!

I'd really appreciate if someone could help me out with this or in case it unrealistic to develop such a rating scale, can someone please explain why?

Thanks a lot.

Sadiq.

In the RF world (my profession), they are a myriad of causes of reduced performance, failure to negotiate properly between client and access device (router), and so on. It's not as simple as wired networks.

Just a big equation with lots of unknowns on both sides.

Today's wireless engineering is all about statistics, due to the way wireless data is framed, distributed in time to minimize impacts of burst errors, suffers multipath which is statistic, and reams of interference types.

Ah, give me a cat5 cable, a coax cable, any time.

 

[dksadiq]

Thanks to you both for your help - it's much appreciated.

I had also began to believe myself that setting a metric wouldn't be feasible.

Tom, as regards the Confidence Interval, I was able to calculate the interval on Excel but then I found it's only a range between which we can be 95% confident that the true mean of the throughput falls - so that wouldn't be of much help here. Thanks for the suggestion though.