Linux Wireless

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.



===== ACS (Automatic Channel Selection) =====

Automatic Channel Selection algorithms and implementations are used to enable interfaces to automatically figure out which channel configuration to use for initiating communication, for any mode of operation which initiates radiation (AP, Mesh, IBSS, P2P). 

There are different algorithms that can be used to finding an ideal channel. We currently only have one, described below. 


===== ACS code =====

All algorithms can be dumped into a generic permissively licensed userspace application which can then be used for integration into different projects. 


===== Survey based algorithm =====

The survey based algorithm relies on the nl80211 survey API command to query the interface for channel active time, channel busy time, channel tx time and noise. The active time consists of the amount of time the radio has spent on the channel. The busy time consists of the amount of time the channel has spent on the channel but noticed the channel was busy and could not initiate communication if it wanted to. The tx time is the amount of time the channel has spent on the channel transmitting data. The survey algorithm relies on these values to build an //interference factor// to give a sense of how much interference was detected on the channel and then picks the channel with the lowest //interference factor//. 

The //interference factor// is defined as the ratio of the observed busy time over the time we spent on the channel, this value is then amplified by the noise floor observed on the channel in comparison to the lowest noise floor observed on the entire band. 

This corresponds to: 


<code>(busy time - tx time) / (active time - tx time) * 2^(chan_nf - band_min_nf)</code>
The coefficient of of 2 reflects the way power in "far-field" radiation decreases as [[linux-wireless@vger.kernel.org|the square of distance from the antenna]] What this does is it decreases the observed busy time ratio if the noise observed was low but increases it if the noise was high, proportionally to the way "far field" radiation changes over distance. Using a minimum noise instead of some "known low noise" value allows this code to be agnostic to any card used, even if the card is yielding incorrect values for noise floor. The minimum noise floor would be the lowest recorded noise floor from all surveyed channels. It is worth showing a few examples of what the multiplier produces to demonstrate how it amplifies the value for high noise. Since we have recorded the lowest observed noise floor by using the delta between what we observed and the lowest observed noise floor on the band we would get multipliers that are always positive and the lowest multiplier would be a value of 1 given that 2^0 = 1. Lets assume the lowest recorded noise floor was -110 dBm here are a few example values: 


<code>mcgrof@tux ~ $ calc
C-style arbitrary precision calculator (version 2.12.3.3)
Calc is open software. For license details type:  help copyright
[Type "exit" to exit, or "help" for help.]

; define f(x) = 2^(x - -110)
f(x) defined
; f(-110)
        1
; f(-109)
        2
; f(-108)
        4</code>
The right hand value of the formula then can be thought of as the amplifier of the observed 'business'. 

If channel busy time is not available the fallback is to use channel rx time. 

Since noise floor is in dBm it is necessary to convert it into Watts so that combined channel intereference (e.g. HT40, which uses two channels) can be calculated easily. 


<code>(busy time - tx time) / (active time - tx time) * 2^(10^(chan_nf/10) + 10^(band_min_nf/10))</code>
However to account for cases where busy/rx time is 0 (channel load is then 0%) channel noise floor signal power is combined into the equation so a channel with lower noise floor is preferred. The equation becomes: 


<code>10^(chan_nf/5) + (busy time - tx time) / (active time - tx time) * 2^(10^(chan_nf/10) + 10^(band_min_nf/10))</code>
This "interference factor" is purely subjective and ony time will tell how usable this is. By using the minimum noise floor we remove any possible issues due to card calibration. The computation of the interference factor then is dependent on what the card itself picks up as the minimum noise, not an actual real possible card noise value. 


===== Total interferece factor =====

The above channel interference factor is calculated with no respect to target operational bandwidth. 

To find an ideal channel the above data is combined by taking into account the target operational bandwidth and selected band. E.g. on 2.4GHz channels overlap w/ 20MHz bandwidth, but there is no overlap for 20MHz bandwidth on 5GHz. 

Each valid and possible channel spec (i.e. channel + width) is taken and its interference factor is computed by summing up intereferences of each channel it overlaps. The one with least total interference is picked up. 

Note: This implies base channel interference factor must be non-negative allowing easy summing up. 


===== Example survet-based ACS analysis printout from hostapd =====


<code>ACS: Trying survey-based ACS
ACS: Survey analysis for channel 1 (2412 MHz)
ACS:  1: min_nf=-113 interference_factor=0.0802469 nf=-113 time=162 busy=0 rx=13
ACS:  2: min_nf=-113 interference_factor=0.0745342 nf=-113 time=161 busy=0 rx=12
ACS:  3: min_nf=-113 interference_factor=0.0679012 nf=-113 time=162 busy=0 rx=11
ACS:  4: min_nf=-113 interference_factor=0.0310559 nf=-113 time=161 busy=0 rx=5
ACS:  5: min_nf=-113 interference_factor=0.0248447 nf=-113 time=161 busy=0 rx=4
ACS:  * interference factor average: 0.0557166
ACS: Survey analysis for channel 2 (2417 MHz)
ACS:  1: min_nf=-113 interference_factor=0.0185185 nf=-113 time=162 busy=0 rx=3
ACS:  2: min_nf=-113 interference_factor=0.0246914 nf=-113 time=162 busy=0 rx=4
ACS:  3: min_nf=-113 interference_factor=0.037037 nf=-113 time=162 busy=0 rx=6
ACS:  4: min_nf=-113 interference_factor=0.149068 nf=-113 time=161 busy=0 rx=24
ACS:  5: min_nf=-113 interference_factor=0.0248447 nf=-113 time=161 busy=0 rx=4
ACS:  * interference factor average: 0.050832
ACS: Survey analysis for channel 3 (2422 MHz)
ACS:  1: min_nf=-113 interference_factor=2.51189e-23 nf=-113 time=162 busy=0 rx=0
ACS:  2: min_nf=-113 interference_factor=0.0185185 nf=-113 time=162 busy=0 rx=3
ACS:  3: min_nf=-113 interference_factor=0.0186335 nf=-113 time=161 busy=0 rx=3
ACS:  4: min_nf=-113 interference_factor=0.0186335 nf=-113 time=161 busy=0 rx=3
ACS:  5: min_nf=-113 interference_factor=0.0186335 nf=-113 time=161 busy=0 rx=3
ACS:  * interference factor average: 0.0148838
ACS: Survey analysis for channel 4 (2427 MHz)
ACS:  1: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=162 busy=0 rx=0
ACS:  2: min_nf=-114 interference_factor=0.0555556 nf=-114 time=162 busy=0 rx=9
ACS:  3: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=161 busy=0 rx=0
ACS:  4: min_nf=-114 interference_factor=0.0186335 nf=-114 time=161 busy=0 rx=3
ACS:  5: min_nf=-114 interference_factor=0.00621118 nf=-114 time=161 busy=0 rx=1
ACS:  * interference factor average: 0.0160801
ACS: Survey analysis for channel 5 (2432 MHz)
ACS:  1: min_nf=-114 interference_factor=0.409938 nf=-113 time=161 busy=0 rx=66
ACS:  2: min_nf=-114 interference_factor=0.0432099 nf=-113 time=162 busy=0 rx=7
ACS:  3: min_nf=-114 interference_factor=0.0124224 nf=-113 time=161 busy=0 rx=2
ACS:  4: min_nf=-114 interference_factor=0.677019 nf=-113 time=161 busy=0 rx=109
ACS:  5: min_nf=-114 interference_factor=0.0186335 nf=-114 time=161 busy=0 rx=3
ACS:  * interference factor average: 0.232244
ACS: Survey analysis for channel 6 (2437 MHz)
ACS:  1: min_nf=-113 interference_factor=0.552795 nf=-113 time=161 busy=0 rx=89
ACS:  2: min_nf=-113 interference_factor=0.0807453 nf=-112 time=161 busy=0 rx=13
ACS:  3: min_nf=-113 interference_factor=0.0310559 nf=-113 time=161 busy=0 rx=5
ACS:  4: min_nf=-113 interference_factor=0.434783 nf=-112 time=161 busy=0 rx=70
ACS:  5: min_nf=-113 interference_factor=0.0621118 nf=-113 time=161 busy=0 rx=10
ACS:  * interference factor average: 0.232298
ACS: Survey analysis for channel 7 (2442 MHz)
ACS:  1: min_nf=-113 interference_factor=0.440994 nf=-112 time=161 busy=0 rx=71
ACS:  2: min_nf=-113 interference_factor=0.385093 nf=-113 time=161 busy=0 rx=62
ACS:  3: min_nf=-113 interference_factor=0.0372671 nf=-113 time=161 busy=0 rx=6
ACS:  4: min_nf=-113 interference_factor=0.0372671 nf=-113 time=161 busy=0 rx=6
ACS:  5: min_nf=-113 interference_factor=0.0745342 nf=-113 time=161 busy=0 rx=12
ACS:  * interference factor average: 0.195031
ACS: Survey analysis for channel 8 (2447 MHz)
ACS:  1: min_nf=-114 interference_factor=0.0496894 nf=-112 time=161 busy=0 rx=8
ACS:  2: min_nf=-114 interference_factor=0.0496894 nf=-114 time=161 busy=0 rx=8
ACS:  3: min_nf=-114 interference_factor=0.0372671 nf=-113 time=161 busy=0 rx=6
ACS:  4: min_nf=-114 interference_factor=0.12963 nf=-113 time=162 busy=0 rx=21
ACS:  5: min_nf=-114 interference_factor=0.166667 nf=-114 time=162 busy=0 rx=27
ACS:  * interference factor average: 0.0865885
ACS: Survey analysis for channel 9 (2452 MHz)
ACS:  1: min_nf=-114 interference_factor=0.0124224 nf=-114 time=161 busy=0 rx=2
ACS:  2: min_nf=-114 interference_factor=0.0310559 nf=-114 time=161 busy=0 rx=5
ACS:  3: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=161 busy=0 rx=0
ACS:  4: min_nf=-114 interference_factor=0.00617284 nf=-114 time=162 busy=0 rx=1
ACS:  5: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=162 busy=0 rx=0
ACS:  * interference factor average: 0.00993022
ACS: Survey analysis for channel 10 (2457 MHz)
ACS:  1: min_nf=-114 interference_factor=0.00621118 nf=-114 time=161 busy=0 rx=1
ACS:  2: min_nf=-114 interference_factor=0.00621118 nf=-114 time=161 busy=0 rx=1
ACS:  3: min_nf=-114 interference_factor=0.00621118 nf=-114 time=161 busy=0 rx=1
ACS:  4: min_nf=-114 interference_factor=0.0493827 nf=-114 time=162 busy=0 rx=8
ACS:  5: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=162 busy=0 rx=0
ACS:  * interference factor average: 0.0136033
ACS: Survey analysis for channel 11 (2462 MHz)
ACS:  1: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=161 busy=0 rx=0
ACS:  2: min_nf=-114 interference_factor=2.51189e-23 nf=-113 time=161 busy=0 rx=0
ACS:  3: min_nf=-114 interference_factor=2.51189e-23 nf=-113 time=161 busy=0 rx=0
ACS:  4: min_nf=-114 interference_factor=0.0432099 nf=-114 time=162 busy=0 rx=7
ACS:  5: min_nf=-114 interference_factor=0.0925926 nf=-114 time=162 busy=0 rx=15
ACS:  * interference factor average: 0.0271605
ACS: Survey analysis for channel 12 (2467 MHz)
ACS:  1: min_nf=-114 interference_factor=0.0621118 nf=-113 time=161 busy=0 rx=10
ACS:  2: min_nf=-114 interference_factor=0.00621118 nf=-114 time=161 busy=0 rx=1
ACS:  3: min_nf=-114 interference_factor=2.51189e-23 nf=-113 time=162 busy=0 rx=0
ACS:  4: min_nf=-114 interference_factor=2.51189e-23 nf=-113 time=162 busy=0 rx=0
ACS:  5: min_nf=-114 interference_factor=0.00617284 nf=-113 time=162 busy=0 rx=1
ACS:  * interference factor average: 0.0148992
ACS: Survey analysis for channel 13 (2472 MHz)
ACS:  1: min_nf=-114 interference_factor=0.0745342 nf=-114 time=161 busy=0 rx=12
ACS:  2: min_nf=-114 interference_factor=0.0555556 nf=-114 time=162 busy=0 rx=9
ACS:  3: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=162 busy=0 rx=0
ACS:  4: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=162 busy=0 rx=0
ACS:  5: min_nf=-114 interference_factor=1.58489e-23 nf=-114 time=162 busy=0 rx=0
ACS:  * interference factor average: 0.0260179
ACS: Survey analysis for selected bandwidth 20MHz
ACS:  * channel 1: total interference = 0.121432
ACS:  * channel 2: total interference = 0.137512
ACS:  * channel 3: total interference = 0.369757
ACS:  * channel 4: total interference = 0.546338
ACS:  * channel 5: total interference = 0.690538
ACS:  * channel 6: total interference = 0.762242
ACS:  * channel 7: total interference = 0.756092
ACS:  * channel 8: total interference = 0.537451
ACS:  * channel 9: total interference = 0.332313
ACS:  * channel 10: total interference = 0.152182
ACS:  * channel 11: total interference = 0.0916111
ACS:  * channel 12: total interference = 0.0816809
ACS:  * channel 13: total interference = 0.0680776
ACS: Ideal channel is 13 (2472 MHz) with total interference factor of 0.0680776</code>

===== Driver implementation details =====

User space can get survey using ''NL80211_CMD_GET_SURVEY'' command. It requires:
  * ''cfg80211'' drivers to implement ''dump_survey'' callback
  * ''mac80211'' drivers to implement ''get_survey'' callback

As explained is algorithm section, driver needs to provide (fill) few different informations about a channel:
  * Noise floor (''NL80211_SURVEY_INFO_NOISE'')
  * Channel time (''NL80211_SURVEY_INFO_CHANNEL_TIME'')
  * Time of channel unavailability which is **one** of:
    * RX time (''NL80211_SURVEY_INFO_CHANNEL_TIME_RX'')
    * Busy time (''NL80211_SURVEY_INFO_CHANNEL_TIME_BUSY'')

===== Hostapd setup =====

Only the following drivers support the current (2014-08-19) survey based ACS implmenetation in hostapd ([[http://w1.fi/cgit/hostap/tree/hostapd/defconfig#n309|http://w1.fi/cgit/hostap/tree/hostapd/defconfig#n309]]). 
  * ath5k 
  * ath9k 
  * ath10k When building make sure you enable ACS in hostapd's .config file: 


<code>CONFIG_ACS=y</code>
To make hostapd do ACS during runtime make sure your hostapd.conf has either: 


<code>channel=0</code>
or 


<code>channel=acs_survey</code>
Hostapd picks a channel automatically. If you don't configure HT40 it will not use HT40. If you do configure HT40 it will use it or fail if it's impossible to setup HT40 (e.g. due regulatory). 


===== Status =====
    * initial implementation is now complete, [[http://marc.info/?l=linux-wireless&m=130870274629048&w=2|ACS hostapd RFC patches]] have been posted. 
    * [[http://marc.info/?l=hostap&m=137569522705154&w=2|improved implementation has been posted]], 
    * As of 31 Aug 2013 ACS has been **merged** into upstream hostapd ([[http://w1.fi/gitweb/gitweb.cgi?p=hostap.git;a=commit;h=50f4f2a066e60be7cc03e291635e9a0f953922bd|final commit]]). 


===== References =====
      * [[http://marc.info/?t=130624150700002&r=1&w=2|Initial ACS algorithm proposal]] 
      * [[http://marc.info/?t=130775164000002&r=1&w=2|ACS code announcement]]
Linux Wireless

User Tools

Site Tools

Page Tools
