Forum

Working for a company that manufactures many battery powered devices, I am interested in good Power Saving methodologies.

I'm not interested in the QOS abilites of WMM per se, but the power save techniques are right on. We'd be willing to install one to get the other if it's good enough.

We know what the WFA says, but that's all hypothetical. Is there anyone out there that can say what their experience with WMM-PS has actually been ?

Alternatively, if you know of any company that currently supports it in a [u]client[/u] device, can you give me their names.

Thanks very much.

Wlanman

Hi.
Yes both our WiFi phones does support WMM-PS. Both the i75 and teh i62.

Talk talk is around 5 times longer with PS.

Martin
Ascom

Martin,

Prior to th WMM-PS, did you have your own propritery PS algorithm? Other than the power savings, how has WMM-PS affected things like Roaming?

I would assume, that being phones, a high priority is low latency. What kinds of tradeoffs did you have to make, or was everything better?

The devices I work with have minimal latency problems. Overall battery life, with occaisional high current draws, and moderate to high amounts of roaming are our main concerns. Except for 1 and 2 Mbps, any other rate woks fine for us.

Hi Wlanman,

I know I'm late by several months to post to this discussion so it may all be for naught, but I just happened to stumble across this thread. Like Martin, I too work for the Wireless Solutions division of Ascom currently and as mentioned we employ WMM-PS at nearly all i62 deployments, and almost all i75 installations (older sites may [i]still[/i] have not updated their WLAN infrastructure to support this). We use this method to great success and you are correct, as the i62 and the i75 are VoWiFi handsets, latency is always of paramount concern.

Obviously if the handsets are left in active mode (i.e. radio is always on), then they will exhibit the least amount of latency and by all benchmarks will have the greatest level of performance, but in doing so will have the lowest possible battery life. To balance this, most VoWiFi vendors will use the benefits of WMM in combination with a DTIM value of 2 for the attributed voice ESSID. This offers an extensive boost to the operating battery life of the handsets while ensuring the radio is listening at a frequent enough interval to best negate the effects of latency.

Speaking from real-world experience, as far as the end user is [i]typically[/i] concerned, there is no tradeoff in audio or roaming performance so long as the DTIM settings are of a value equal to or less than 5. Ascom has found that a DTIM value of 5 to be the highest recommended value that offers the greatest benefit to the battery life of our devices, yet a negligible to non-existent impact on the quality of voice calls and handover behavior. I say typically because as we all know, no WLAN environment is ever the same and will often require different client parameters and configurations.

Also keep in mind that all of this is based upon our advanced WLAN chipsets, as well as the fact that the enterprise grade WLAN infrastructures in question were properly surveyed and designed/redesigned for VoWiFi applications, so the APs/WLCs are optimally placed and configured for roaming clients. If you are not operating in a voice-grade WLAN, then be aware that having a radio silent for too long a period of time between active states may likely introduce idiosyncrasies (particularly in roaming conditions) that you will have to both test and account for. However, my experience in optimized and even rudimentary WLAN environments has always seen WMM-PS to almost always be of great benefit to applicable end users if implemented properly.

As to your other question of whether or not Ascom had a proprietary power save algorithm prior to the introduction or widespread adoption of WMM-PS; the answer is yes... :)

Michael,

Thank you very much for your response. Information with this amount of detail has been hard to find, and is always valuable.

Since my original post, I have found details on two other VoWiFi providers' hardware. Both of them insist on a DTIM count of one. It makes sense to me that a value of two would produce good results for both PS and roaming. It's also impressive.

I know that DTIMS of 5 already cause us problems, and unfortunately some customers use a setting of 10 - which I believe is the default on some Symbol AP's. Sometimes it is hard to convince our customers that low DTIM numbers also reduce Roaming problems.

As for other devices, besides VoWiFi, that use WMM-PS I have not had much luck getting any details. Yes, I can go on the WFA website and search for WMM-PS devices, but it still does not help me much prove or disprove the utility of WMM-PS. I think a DTIM count of 2 might be a good value to experiment with.

I am especially glad to hear that you are happy with WMM-PS, even though you had a prior PS solution. It obvioulsy must be worth the cost of certification.