11-14/0523r0

April 2014
doc.: IEEE 802.11-14/0523r0
MAC simulation results for Dynamic
sensitivity control (DSC - CCA adaptation)
and transmit power control (TPC)
Date: 2014-04-17
Authors:
Name
Company
Address
Imad Jamil
Orange
Laurent Cariou
Orange
Thomas Derham
Orange
Jean-Pierre Le Rouzic
Orange
4 rue du clos Courtel
35512 Cesson Sévigné
France
4 rue du clos Courtel
35512 Cesson Sévigné
France
9F Keio Shinjuku
Oiwake Bldg. Shinjuku
3-1-13, Tokyo, Japan
4 rue du Clos Courtel
35512 Cesson Sevigne
France
Submission
Phone
Slide 1
email
[email protected]
+33 299124350
[email protected]
+81 3 5312
8563
[email protected]
+33 299124893
[email protected]
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Context
• In dense environments, CSMA-CA parameters (especially CCA
physical carrier sensing) as defined in the standard are quite
conservative.
– reducing reuse between neighboring cells
• Several presentations have takled this problem:
– adaptation of this CCA (dynamic sensitivity control)
– adaptation of transmit power
• In this presentation, we run a set of simulations to confirm/infirm
first insights regarding these schemes
– MAC system simulator with simple PHY abstraction
Submission
Slide 2
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
co-channel interference (CCI) from
neighboring cells
Useful Rx power
AP
Min SINR to receive MCSx
(sensitivity)
STA
CCI
Co-Channel interference (CCI)
CCA
CCI
Noise floor
Interfering AP
Simultaneous transmission
STA
•
In scenarios where the useful receive power is always sufficiently higher than
interference, reuse between neighboring cells is possible
–
but currently prevented by CCA threshold
Submission
Slide 3
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
How to enable reuse
1: Transmit power control
2: CCA control (DSC)
AP
AP
STA
STA
Co-Channel interference (CCI)
Co-Channel interference (CCI)
Interfering AP
STA
Interfering AP
Simultaneous
STA transmission
Simultaneous
transmission
Useful Rx Power
CCI
CCA
Submission
Useful Rx Power
Min SINR to receive MCSx
(sensitivity)
Min SINR to receive MCSx
(sensitivity)
CCI
CCI
CCA
Noise floor
Slide 4
Noise floor
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Simulate simple algorithms
• DSC (CCA control):
– each STA adjust CCA to CCA = Useful Rx Power – Margin
AP
STA
tune CCA level
tune CCA level (and receiver sensitivity)
• TP control:
•
each receiver requests transmitter to adjust TP so that it receives the Useful Rx Power at
Margin dB above classical CCA (-82dBm)
AP
tune TP
STA
• In each case, the margin has a strong relationship with the min SINR
experienced by STAs
– we are not discussing any protocol in here
Submission
Slide 5
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Simulation scenario
BSS
Cluster
3
Submission
Slide 6
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Simulation scenario
Close to scenario 3
- First tier only (7 BSSs), 8 STAs per BSS, Single channel for reuse 3
Tx : 15 dBm
[21 m]
[7 m]
Tx :15 dBm
Submission
Slide 7
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Simulation scenario
Parameters
PHY
oIEEE 802.11n
oPath Loss: ITU UMi (23.3+36.7log10(d)+ 21log10(2400/900MHz))
oBand: 5 Ghz
oChannel: 20 MHz
oTx power: 15 dBm
Traffic
oFull buffer UDP traffic
oDL (AP->STA)
oUL (STA->AP)
Simple metric
Aggregate throughput
Fixed MCS or rate adaptation AARF
Submission
Slide 8
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Fixed MCS7 – DSC and TPC
Different margin (10-15-20-25-30-35-40-50)
No DSC
Submission
No TPC
DSC
Slide 9
TPC
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Rate control – DSC and TPC
Different set of MCSs for rate control:
all MCSs, MCS 2-7, MCS 3-7, MCS 4-7, MCS 5-7, MCS 6-7
No DSC
Submission
No TPC
DSC
Slide 10
TPC
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
First observations
• CCA and TPC are strongly increasing reuse and
aggregate throughput
– Margin optimization
• Weak rate control algorithms make aggregate
throughput collapse
• The margin gives an indication on min SINR
– if used for suppression of MCS usage below this min SINR
– it leads to strong improvement of rate control efficiency and to
aggregate throughput gains
Submission
Slide 11
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Impact of legacy devices
• mix of legacy (not implementing DSC or TPC) and DSC/TPCcapable STAs (implementing DSC or TPC)
– 1 legacy STA per BSS
Submission
Slide 12
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Rate control – DSC– mix with legacy devices
All DSC-capable STAs
No DSC
Submission
DSC-capable STAs + 7 legacy STAs
DSC
No DSC
Slide 13
DSC
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Rate control - TPC – mix with legacy devices
All TPC-capable STAs
No TPC
Submission
TPC-capable STAs + 7 legacy STAs
TPC
No TPC
Slide 14
TPC
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
First observations for mix with legacy
• As expected, the impact of legacy devices on aggregate
throughput is way stronger with TPC than with DSC
– with TPC, those legacy are disabling any reuse possibility
– with TPC, those legacy are also disturbing the relationship between the margin and
the min SINR and cancels MCS suppression benefits in rate control algorithms
Submission
Slide 15
Imad Jamil (Orange)
April 2014
doc.: IEEE 802.11-14/0523r0
Conclusion/next steps
• Preliminary results with MAC system simulator are
showing
– very good per user throughput gains both for TPC and DSC
– additional benefit of the « margin » for rate adaptation
improvement through desabling the use of low MCSs
– strong impact of legacy devices on TPC reuse efficiency
• Next steps
–
–
–
–
Submission
simulator evolution (calibrated)
different scenarios
analysis of potential starvation of legacy devices in case of DSC
analysis of RTS/CTS impact
Slide 16
Imad Jamil (Orange)