Ping of AP from Laptop slow on battery

Hi,

I have an HP Laptop (Compaq nc6120) which
has integral Intel Centrino wireless and I also
have a Cisco 857W Router.

OS is Windows XP SP2.
I am using WPA TKIP.

On one occasion I noticed that the wireless performance was poor and a
set the dot11 interface as follows:-
speed basic-1.0 basic-2.0 basic-5.5 6.0 9.0 basic-11.0 basic-12.0
basic-18.0 basic-24.0

The PC is 10 feet from the AP, line of sight is unobstructed.

The issue:-

I have now noticed that the ping time to the router over the
wireless link was sometimes concerningly slow.

I have now noticed that the different behaviours
occur depending on whether the Laptop is connected to the
mains or running off battery.

Here is what I see:-
The #### indicate that the power lead was removed or
re-connected. I flagged a couple of interesting reports with #.

Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255 ####
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=12ms TTL=255
Reply from 172.18.46.1: bytes=32 time=16ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255 #
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=17ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255 #
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=18ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255 ####
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255
Reply from 172.18.46.1: bytes=32 time=1ms TTL=255

I have tested pinging to 127.0.0.1 and the address of
the PC's wireless card 172.18.46.201.

In these cases there is no change to the reported RTT
which is <1ms so I don' think that the CPU is
slowing down such that the issue arrises.

I have had a look in Windows/ControlPanel/Power
however there is nothing thre about wireless.

Can anyone explain what might be going on here?

Thanks.

(E-Mail Removed) wrote:
>I have now noticed that the different behaviours
>occur depending on whether the Laptop is connected to the
>mains or running off battery.

>I have had a look in Windows/ControlPanel/Power
>however there is nothing thre about wireless.

Check the WiFi card in the Device Manager for a power management tab,
or something in details for power use...

>anybod...@hotmail.com wrote:
>>I have now noticed that the different behaviours
>>occur depending on whether the Laptop is connected to the
>>mains or running off battery.
>>I have had a look in Windows/ControlPanel/Power
>>however there is nothing thre about wireless.

William P. N. Smith wrote:-
>Check the WiFi card in the Device Manager for a power management tab,
>or something in details for power use...

Thanks. There were settings there.

Still don't understand it though. Will check out Intel.

Interestingly the phenomenon seems to have gone away
while I am doing a 100M download from the internet.

In article <(E-Mail Removed) .com>,
<(E-Mail Removed)> wrote:
>
>>anybod...@hotmail.com wrote:
>>>I have now noticed that the different behaviours
>>>occur depending on whether the Laptop is connected to the
>>>mains or running off battery.
>>>I have had a look in Windows/ControlPanel/Power
>>>however there is nothing thre about wireless.
>
>William P. N. Smith wrote:-
>>Check the WiFi card in the Device Manager for a power management tab,
>>or something in details for power use...
>
>Thanks. There were settings there.
>
>Still don't understand it though. Will check out Intel.
>
>Interestingly the phenomenon seems to have gone away
>while I am doing a 100M download from the internet.


802.11's power-saving mode works by having the mobile station's receiver
stay in a shutdown / "sleep" state during idle periods, waking up only
at set intervals (typically every 10-20ms) to check for an incoming
packet. Its wake interval is synchronized with the access point, which
holds any packets destined for a sleeping client until the next wake-up
cycle.

Once a steady data transfer is underway, the mobile receiver will remain
powered on full-time until the link becomes idle idle again (for a
configurable period of time, e.g. 100ms). So, bulk data tranfers aren't
affected much, but interactive/sporadic communication sees higher
latency, like you noticed with the ping test.

I tried this for a while with my laptops, but thought the battery life
savings, at less than 10%, wasn't worth the extra lag in ssh & telnet
session. PDAs, with their lower power budget proably see more of a
benefit...

It sounds like your laptop enables Wifi power-saving automatically,
whenever it's on battery, but this should be overridable in the Control
Panel settings.

--
Jordan.

Jordan said:-

>802.11's power-saving mode works by .....


Thanks, I guess that I had just about enough experimental
data to work that out for myself however I had
never heard of a power saving mode.

Any recommended books on 802.11 that actually tell
you how it works?

(E-Mail Removed) (Jordan Hazen) hath wroth:

>802.11's power-saving mode works by having the mobile station's receiver
>stay in a shutdown / "sleep" state during idle periods, waking up only
>at set intervals (typically every 10-20ms) to check for an incoming
>packet. Its wake interval is synchronized with the access point, which
>holds any packets destined for a sleeping client until the next wake-up
>cycle.

A bit more detail on how the power save mode works:
| http://www.tkn.tu-berlin.de/publicat...eletter.fm.pdf
| http://www.wi-fiplanet.com/tutorials...le.php/1015781
In infrastructure mode, power saving is coordinated with the beacon
interval. I've found that it sometimes fails to function if one turns
off SSID broadcast in the wireless access point. Enabling flow
control also does weird things. Some access points have a minimal
packet buffer, which promptly overflows when a clients power save bit
is set. The AP has to buffer packets while the client is dozing. When
I set my BEFW11S4 wireless router to beacon once per second instead of
the usual 10 times per second, large file downloads began slowing down
or hanging as the card (WG511) would go to doze mode between beacons.
Lots of ways to do it wrong or screw it up.

>Once a steady data transfer is underway, the mobile receiver will remain
>powered on full-time until the link becomes idle idle again (for a
>configurable period of time, e.g. 100ms). So, bulk data tranfers aren't
>affected much, but interactive/sporadic communication sees higher
>latency, like you noticed with the ping test.

From:
| http://www.wi-fiplanet.com/tutorials...le.php/1015781
Decrease in throughput. Keep in mind that to achieve significant
battery savings using sleep mode, you have to be willing to live
with extremely low throughput. Some applications that require
frequent communications with the clients will not operate well
with power saving turned on.

Methinks the above is generally correct. One would logically expect
the power save mode to be disabled temporarily for large file
transfers. I don't think all cards do this.

I know that Intel 2200BG with Proset 10.1 does it right. If you sniff
the wireless traffic, the power save bit in the management frames is
reset to 0 (off) when traffic gets heavy, and returns back to 1 (power
save on) when things slow down. I vaguely recall the response time is
about 4 seconds.

>I tried this for a while with my laptops, but thought the battery life
>savings, at less than 10%, wasn't worth the extra lag in ssh & telnet
>session. PDAs, with their lower power budget proably see more of a
>benefit...
>
>It sounds like your laptop enables Wifi power-saving automatically,
>whenever it's on battery, but this should be overridable in the Control
>Panel settings.

--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558