Power Save mode

[ccie]

The 802.11 standard specifies use of PS-Poll frames for use in Power Save mode. The Cisco 802.11b and 802.11a cards operate exactly according to the standard, which is good. Once they enter Power Save mode, they stay in Power Save mode, but look at the TIM field in the Beacons to decide whether or not to sent PS-Poll frames to the AP after waking up. The Cisco 802.11a/b/g card (CB21AG) does not do this the same way. Instead, when it powers up, it immediately sends Null Function Data frames to the AP with the Power Mgmt bit in the Frame Control field of the MAC header set to 0. It does so without even looking at the TIM field in the first Beacon it receives after waking up. Setting this bit to 0 tells the AP to take it out of Power Save mode and to send any/all queued traffic. When the last queued frame is sent to the station (with the More Data bit set to 0), the station sends another Null Function Data frame with the Power Mgmt bit set to 1 (meaning "I'm going back to sleep"). This functionality is not according to the standard. Why did Cisco previously comply with the standard and now it does not?

[dixho]

I think that this question has been raised by many customers. A couple of bugs has been opened for this: CSCee90576 and CSCee90262. CSCee90262 is a duplicate of CSCee90576. CSCee90576 is junked by the software development group. The comment from the DE (Development Engineer) is as follows:

"Reason for this action:This is the expected behavior of the client. The client does

a "fake" sleep so that the AP will save all packets for the STA when it is off channel doing a

background scan or a Radio Measurment background scan. This is an acceptable method of active

scanning when away from your channels according to the 802.11 specification."