Fair Real-time Traffic Scheduling over A Wireless Local Area Network

23
Scheduling over A Wireless Local Area Network Maria Adamou, Sanjeev Khanna, Insup Lee, Insik Shin, and Shiyu Zhou Dept. of Computer & Information Science University of Pennsylvania, USA

description

Fair Real-time Traffic Scheduling over A Wireless Local Area Network. M aria Adamou, S anjeev Khanna, I nsup Lee, Insik Shin, and S hiyu Zhou Dept. of Computer & Information Science University of Pennsylvania , USA. Real- t ime Communication over Wireless LAN. MH 1. BS. MH 2. - PowerPoint PPT Presentation

Transcript of Fair Real-time Traffic Scheduling over A Wireless Local Area Network

Page 1: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

Fair Real-time Traffic Scheduling

over A Wireless Local Area Network

Maria Adamou, Sanjeev Khanna,

Insup Lee, Insik Shin, and Shiyu Zhou

Dept. of Computer & Information Science

University of Pennsylvania, USA

Page 2: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

2

Real-time Communication over Wireless LAN

BSMH1

MH3

MH2

Page 3: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

3

Wireless LAN MAC Protocol IEEE 802.11 – standard

DCF (distributed) Contention-based transmission

PCF (centralized)

Contention-free (CF) transmission BS schedules CF transmissions by polling

Page 4: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

4

Wireless Network Characteristics Unpredictable Channel Error

location dependent bursty

BSMH1

MH3MH2

Page 5: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

5

Challenges How do channel errors affect real-time

transmissions? QoS degradation Wireless channel error model

How does BS schedule real-time transmissions with unpredictable errors? Real-time scheduling objective considering

QoS degradation with errors Real-time scheduling algorithm

Page 6: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

6

Outlines Real-time traffic model Scheduling objectives Theoretical results Online scheduling algorithms Simulation results Conclusion

Page 7: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

7

Real-time Traffic Model Periodic packet generation (release time) Soft deadline

Upon missing deadline, a packet is dropped

Acceptable packet loss (deadline miss) rate Degradation = actual loss rate – acceptable loss

rate

The same packet length (execution time)

Page 8: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

8

Scheduling objectives1. Fairness (considering each flow)

Location dependent channel errors Minimizing the maximum degradation

2. Throughput (considering the system) Maximizing the overall system throughput

(fraction of packets meeting deadlines)

Online scheduling algorithm without knowledge of error in advance

Page 9: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

9

Theoretical results No online optimal algorithm

Performance ratio of an online algorithm w.r.t. optimal for throughput maximization, two for achieving fairness, unbounded For the combined objectives, unbounded

A polynomial time offline algorithm that optimally achieves our scheduling objectives

Page 10: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

10

Online scheduling algorithms EDF (Earliest Deadline First)

GDF (Greatest Degradation First)

EOG (EDF or GDF)

LFF (Lagging Flows First)

Page 11: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

11

EDF (Earliest Deadline First)when a new packet is available

3 0.2

Di εi

4 0.4 3 0.3 1 0.1

EDF QueueScheduler

when it dispatches

Page 12: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

12

GDF (Greatest Degradation First)when a new packet is available

3 0.2

Di εi

1 0.1 3 0.3 4 0.4

GDF QueueScheduler

when it dispatches

Page 13: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

13

EOG (EDF or GDF)when a new packet is available

3 0.2

4 0.4 3 0.3 1 0.1

EDF Queue

Scheduler

when it dispatches

1 0.1 3 0.3 4 0.4

GDF Queue

If there is a packet that will miss its deadline after next slot

Otherwise

Page 14: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

14

LFF (Lagging Flows First)when a new packet is available

3 0.2

Di εi

4 0.4

LFF Array

4

index

2

1 0.11

3 0.33

Page 15: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

15

LFF (Lagging Flows First)when a new packet is available

3 0.2

Di εi

4 0.4

LFF ArrayScheduler

when it dispatches

4

index

2

1 0.11

3 0.33

3 0.2

Page 16: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

16

LFF (Lagging Flows First)when a new packet is available

3 0.2

4 0.4 2 0.3 1 0.1

EDF Queue

Scheduler

when it dispatches

1 0.1 2 0.3 4 0.4

GDF Queue

If there is a packet that will miss its deadline after next slot

Otherwise

Page 17: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

17

Simulation – Performance Metrics

1. Degradation (for each flow) Fraction of packets lost beyond the

acceptable packet loss rate

2. Throughput (over all flows) Fraction of successfully transmitted

packets

Page 18: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

18

Simulation – Error Modeling Random blackouts (wi) for error period

Error duration rate =

BSMH1

MH3MH2

MH1

tmaxt0

MH2 MH3

wi

maxt

wi

i

Page 19: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

19

Results – Max Degradation

0

0.1

0.2

0.3

0 0.1 0.2 0.3 0.4Error Duration Rate

Deg

rada

tion

deg

ree

EDF

GDF

EOG

LFF

Page 20: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

20

Results – Throughput Ratio

0.98

0.985

0.99

0.995

1

1.005

1.01

1.015

1.02

0 0.1 0.2 0.3 0.4Error Duration Rate

Thr

ough

put r

atio

vs

EO

G

EDFGDFEOGLFF

Page 21: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

21

Related Work QoS guarantees over wireless links

No consideration of fairness issue WFQ over wireless networks

No consideration of deadline constraint QoS degradation considering deadline

Imprecise computation IRIS (Increased Reward with Increased Service) (m,k)-firm deadline model DWCS (Dynamic Window-Constrained Scheduling)

Page 22: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

22

Conclusion Scheduling objectives

1. Fairness – minimizing the maximum degradation

2. Overall throughput maximization

Theoretical results No online algorithm can be guaranteed to

achieve a bounded performance ratio for the scheduling objective

Page 23: Fair Real-time Traffic Scheduling  over  A  Wireless Local Area Network

23

Conclusion Online algorithms

For fairness objective1. LFF 2. GDF 3. EOG 4.EDF

For maximum throughput objective1. EDF 2. LFF3. EOG 4.GDF

Future work Variable length packets Other measures of fairness