Spatial Backoff Contention Resolution for Wireless Networks

1
Spatial Backoff Contention Resolution for
Wireless Networks

• Xue Yang and Nitin H. Vaidya
• University of Illinois at Urbana-Champaign

2
Transmissions Compete for Space
Space occupied by A or B
A
B
3
Temporal Contention Resolution
Space
Time

• Temporal contention resolution resolves the local
  channel contention in time dimension

4
Transmissions Compete for space
A
B
5
Spatial Contention Resolution
Space
Time
B transmits
Space
A transmits
Time
6
Hybrid Contention Resolution
S1
?
S4
S6
S5
S3
S2
7
Spatial Contention Resolution

• Control the contending area by adapting
  transmission power/rate/carrier-sensing
• Goal
• In this work
• Joint adaptation of transmission rate and
  carrier-sense (CS) threshold, assuming a fixed
  transmission power

Contending area Desired contention level
8
How CS Threshold Controls Contending Area
B
D
C
A
F
E
Signal Strength
CS Threshold
distance
9
How CS Threshold Controls Contending Area

• Larger CS threshold leads to smaller contending
  area

B
D
C
A
F
E
Signal Strength
CS Threshold
distance
10
Transmission rate needs to be adjusted

• Larger CS threshold leads to higher interference
• Transmission rate depends on Signal-to-Interferenc
  e-Noise Ratio

B
D
C
A
F
E
Signal Strength
CS Threshold
distance
11
Why Smaller Contending Region?

• Reduce the collision probability due to local
  contention
• Smaller number of locally competing nodes
• Reduce the rate-independent MAC overhead
• Channel time consumed by physical layer
  convergence procedure (PLCP) preamble and header
  (192 ms in IEEE802.11b).
• Inter-frame space (DIFS, SIFS, etc) and the
  backoff durations
• A smaller fraction of channel capacity is wasted
  in MAC overhead, if the channel operates at a
  lower rate.

12
Optimal CS threshold and transmission rate depend
on transmission density
13
Local information leads to sub-optimal decisions

• Using cs1, node 1 and 2 transmit alternately with
  rate R1
• Using cs2, node 1 and 2 transmit concurrently
  with rate R2
• Optimal CS threshold for node 1 is cs1 if R2ltR1/2

• Using cs2, node 1 transmits with rate R2, node 2
  transmits concurrently with rate R3
• Optimal CS threshold for node 1 is cs2 if R2R3gtR1

14
Dynamic Spatial Backoff Algorithm

• To search the two-dimensional space defined by
  the CS threshold and the transmission rate.
• K discrete rates RateiltRatej, if iltj,
  i,j?(1, K)
• Smallest CS threshold CSi for Ratei
• CSigtCSj, if iltj.

• To search the two-dimensional space defined by
  the CS threshold and the transmission rate.
• K discrete rates RateiltRatej, if iltj,
  i,j?(1, K)
• Smallest CS threshold CSi for Ratei
• CSigtCSj, if iltj.

Reduce the search space to the subspace above or
on the diagonal line
15
Search Rules

• Rule 1 when transmission are successful
• Increase transmission rate by one level
• Update the CS threshold for the new rate with the
  CS threshold associated with the old rate

16
Search Rules

• Rule 2 when transmission fail and the operating
  point is above the diagonal line
• Keep the transmission rate
• Decrease CS threshold by one level

17
Search Rules

• Rule 3 when transmission fail and the operating
  point is on the diagonal line
• Decrease transmission rate by one level
• Apply the CS threshold previously associated with
  the new rate

18
Search Rules

• Rule 4 To avoid starvation when probing small
  threshold, if no transmission attempts in a
  certain time period
• Reduce transmission rate by one level
• Apply the CS threshold previously associated with
  the new rate

19
Stability Properties
20
(No Transcript)
21
Open Problems

• Jointly transmission power, CS threshold, and
  transmission rate adaptation
• Search space
• Search rules
• Integration of temporal and spatial contention
  resolutions
• Impact of traffic and channel variations