Submission Page 1 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Radio Measurement Actions Daryl Kaiser (Cisco Systems) 12 November 2002

Submission Page 2 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Where do radio measurement actions fit into the proposed RRM framework? First, Define and Add Passively Observable Parameters to MIB Standard Impact Minimal While the above is being balloted Transfer of Parameters between STA to AP Actively Measured Parameters: extend TGh? –May Need some architectural thinking While the above is being balloted Work on Control Issues: Maybe a next TG –Definitely Need some architectural thinking

Submission Page 3 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Defining the focus of this presentation What it addresses: Measurements that require specific action Measurements that are not part of normal processing, or Measurements that make sense to perform in “snapshots” Suggests measurement contents for further discussion Not addressed: Status and statistics requiring no specific action E.g., authentication/association status, link counter statistics Can populate MIB objects for these without measurement action

Submission Page 4 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Specific RRM goals that require radio measurement action Detect possible rogue APs to help the WLAN administrator plug security holes Quantify WLAN radio topology to facilitate automatic frequency selection and AP/client power limits Measure BSS overlap to help balance coverage, capacity and QoS (current topic in TGe) Quantify each station’s local performance for better network visualization and to facilitate roaming and load balancing Detect non interference and quantify noise to facilitate adjustments in WLAN configuration

Submission Page 5 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Requests Set through measurement action table in MIB Include any required measurement parameters May pertain to AP radio and/or client radios Measurements Serving channel is measured without disrupting traffic Off-channel measurements temporarily disrupt traffic Station’s traffic is buffered if it measures off-channel Station may decline if measurement poses undue burden Reports Reports communicated through new MIB tables Define a request/report mechanism

Submission Page 6 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Action #1 = Beacon Report (and probe responses, too) Purpose: Detects possible rogue APs Quantifies WLAN radio topology Measures BSS overlap How it works: Each request specifies channels or asks for current beacon table Measure at AP radio and/or client radios Measure on serving channel or off-channel For each detected beacon and probe response, –Always report channel, BSSID, RSSI and time of detection –Optionally report other fixed fields and information elements

Submission Page 7 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Diagram of Beacon Request/Report Access Point Client Station Write beacon reports to MIB Detected BSSID, SSID, channel Other detected beacon parameters TSFTIMER offset versus serving AP Detected signal strength per STA One request Periodic poll/trap Periodic requests One report per request Retrieve reports Poll or trap (TBD) External Network Manager Schedule radio measurements AP follows scan parameters Trigger measurements on schedule Measure locally and request of clients Request via MIB Scan interval and channels Active and/or passive scan AP and/or client stations Send one report per request Passive co-channel scan Active off-channel scan Accumulated beacon table MIB Local measurement access Client saves in MIB Available via MLME-SAP Need SME vendor support Issues How should this work in IBSS? When is STA allowed to decline? How to enforce minimal disruption?

Submission Page 8 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Action #2 = Extended CCA Report Purpose: Quantifies each station’s current local performance How it works: Measure at AP radio and/or client radios Measure on serving channel only Measured parameters –CCA busy fraction indicates degree of local contention at STA –Histogram of recent backoff translates to radio-related jitter at STA

Submission Page 9 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Action #3 = Non RPI Histogram Report Purpose: Detects non interference and quantifies noise How it works: Measure at AP radio and/or client radios Measure on serving channel only Measured parameters –Received power indicator (RPI) is a quantized measure of received power –TGh defines eight RPI levels with 5-dB ranges for inner levels –RPI Histogram measures occurrences of each RPI level –RPI Histogram in TGh does not distinguish non energy

Submission Page 10 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems What could external management entity do if AP detects acceptable CCA, but station detects unacceptably high CCA ? With just CCA information…not much could be done Need to know which stations are causing the contention Define one more radio measurement action for this case: Action #4 = Frame Report

Submission Page 11 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Action #4 = Frame Report Purpose: Identifies a station’s local contending stations How it works: Measure at AP radio or client radios, but typically client Measure on serving channel only Report one sub-element for each detected contending station – Transmitter MAC address – BSSID to which contending STA belongs – Number of frames decoded from STA – Average signal strength of frames from STA

Submission Page 12 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Store the reported measurement table in MIB under SMT dot11reportedStatsTable Measurements taken at STA itself Measurements taken by other STAs and reported back

Submission Page 13 November 2002 doc.: IEEE /677r0 Daryl Kaiser, Cisco Systems Summary of proposed radio measurement actions Beacon report detects possible rogue APs, quantifies WLAN radio topology and measures BSS overlap Extended CCA report quantifies each station’s current local performance in terms of CCA busy fraction and backoff experience Non RPI histogram report detects non interference and quantifies noise Frame report identifies a station’s local contending stations