Energy Aware Scheduling and Queue Management for Next Generation - - PowerPoint PPT Presentation

energy aware scheduling and queue management for next
SMART_READER_LITE
LIVE PREVIEW

Energy Aware Scheduling and Queue Management for Next Generation - - PowerPoint PPT Presentation

Energy Aware Scheduling and Queue Management for Next Generation Wi-Fi Routers Husnu S aner Narman Mohammed Atiquzzaman School of Computer Science University of Oklahoma, USA. husnu@ou.edu http://students.ou.edu/N/Husnu.S.Narman-1 March


slide-1
SLIDE 1

Energy Aware Scheduling and Queue Management for Next Generation Wi-Fi Routers

Husnu Saner Narman Mohammed Atiquzzaman

School of Computer Science University of Oklahoma, USA. husnu@ou.edu http://students.ou.edu/N/Husnu.S.Narman-1 March WCNC-NGWiFi 2015

slide-2
SLIDE 2

Outline

  • Introduction
  • Multi Band Router Architecture
  • Energy Aware Algorithm
  • Results
  • Conclusion

Husnu S. Narman

slide-3
SLIDE 3

2.4 GHz 11 Mbps DSSS

Wi-Fi (802.11) Over Generation

Husnu S. Narman 3 2.4 GHz 1, 2 Mbps DSSS,FHSS b 5 GHz 54 Mbps OFDM 2.4 GHz 54 Mbps DSSS,OFDM a 97 g 2.4, 5 GHz 65, 135 Mbps OFDM n 5 GHz 780 Mbps OFDM ac 2.4, 5, 60 GHz 7 Gbps OFDM ad

Multi-band

Future Ones ah: 0.9 GHz aj: 45, 60 GHz ax: 2.4, 5 GHz

slide-4
SLIDE 4

Current Multi Band (802.11n generation)

Husnu S. Narman 4

1st Band: 2.4 GHz 2nd Band: 5 GHz The benefit of using multi-band router is less interference, higher capacity and better reliability.

Conclusion Energy Multiband Result Introduction

slide-5
SLIDE 5

Current Multi-Band Queuing System

Husnu S. Narman 5

𝑅𝑐

Band-b

𝑅𝑑

Band-c

𝑅𝑏

Band-a Conclusion Energy Multiband Result Introduction

slide-6
SLIDE 6

Multi-shared-band Router (Suggested)

Husnu S. Narman 6

2.4 GHz 5 GHz 2.4 GHz 5 GHz

Conclusion Energy Multiband Result Introduction

slide-7
SLIDE 7

Multi-shared-band Queuing System

Husnu S. Narman 7

𝑅𝑐

Band-b

𝑅𝑑

Band-c

𝑅𝑏

Band-a Conclusion Energy Multiband Result Introduction

slide-8
SLIDE 8

Current Multi-band and Multi-shared-band Routers

  • Energy consumption is high

– $27 per year for a router even for stand by (Ecotricity)

Husnu S. Narman 8 Conclusion Wi-Fi

  • bjective

LTE-A Introduction

slide-9
SLIDE 9

Decreasing energy consumption of multi-band routers.

Husnu S. Narman 10 Conclusion Energy Multiband Result Introduction

Objective

slide-10
SLIDE 10

Current Solution for Energy Awareness

  • Channel selection based on power consumption by considering

QoS.

Husnu S. Narman 13

2.45-2.55 GHz: 5 dB 5.65-5.75 GHz: 10 dB

  • User has Antenna Selection in MIMO

– Good for transferring data but not for stand by.

  • Using small packets

Conclusion Energy Multiband Result Introduction

slide-11
SLIDE 11

Proposed: Energy Aware Scheduling Algorithm and Queue Management (e-ASA)

Husnu S. Narman 15

  • Follows sleep and wake-up procedure of bands
  • Depends on incoming traffic rates and QoS

Conclusion Energy Multiband Result Introduction

slide-12
SLIDE 12

e-ASA Based Multi-shared-band Queuing System

Husnu S. Narman 16

𝑅𝑐

Band-b

𝑅𝑑

Band-c

𝑅𝑏

Band-a Conclusion Energy Multiband Result Introduction

𝜈𝑏 𝜈𝑐 𝜈𝑑

slide-13
SLIDE 13

e-ASA Based Multi-shared-band Scheduling

Husnu S. Narman 17 Conclusion Energy Multiband Result Introduction

Traffic Arrival and Packet transfer rate is measured The required number of band and buffer size is determined Based on active and non active bands, the communication links are created. To increase QoS, first send packets over Least Loaded and highest CQI for each device

𝜇 𝑏𝑜𝑒 𝜈 𝐸 = 𝜍𝑂 1 − 𝜍 1 − 𝜍𝑂+1 𝛿 = 1 − 𝐸 Expected throughput rate

  • Coverage
  • Exp. throughput
  • Packet Trans.
slide-14
SLIDE 14

Energy Modeling

Husnu S. Narman 18 Conclusion Energy Multiband Result Introduction

  • Energy modelling for only downloading
  • Depends on idle or busyness of bands

𝐹𝑏 = 𝑈 ∗ 𝛽 ∗ 𝜍𝑏 + 𝛾 ∗ (1 − 𝜍𝑏 )

Simulation time Energy consumption during data transfer time Band is busy Energy consumption during idle time Band is idle

𝐹 = 𝐹𝑏 + 𝐹𝑐 + 𝐹𝑑

Total energy consumption

slide-15
SLIDE 15

Simulation Environments

Husnu S. Narman 19 Conclusion Energy Multiband Result Introduction

slide-16
SLIDE 16
  • Discrete event simulation by following M/M/3/N.
  • 100 realizations for different number of users with

increasing data traffic.

  • We compare

– Single (Single band router), – Current (Current multi band router), – Shared (Multi-shared-band router) – Shared (e-ASA ) (Multi-shared-band with energy aware scheduling algorithm)

Results

Husnu S. Narman 20 Conclusion Energy Multiband Result Introduction

slide-17
SLIDE 17

Band Usage

Husnu S. Narman 21 Band usages of multi-band routers are lower. Objective Observing effects of number

  • f users on band usage of all

bands. Band Usages of Single and e-ASA are same and lower than multi-band routers. Band Usages of e-ASA is changeable because of dynamic allocation. Conclusion Energy Multiband Result Introduction

slide-18
SLIDE 18

Throughput Rate

Husnu S. Narman 22 e-ASA does not decrease the throughput rate comparing to multi-shared-band. Objective Observing effects of number

  • f users on throughput rate.

Conclusion Energy Multiband Result Introduction Throughput rate of single are lower. Throughput rate of e-ASA and shared are higher than

  • ther cases.
slide-19
SLIDE 19

Energy Consumption

Husnu S. Narman 23 Energy consumption of e-ASA is lower than

  • ther multi-band routers except when three

bands are fully used. Objective Observing effects of number

  • f users on energy usage.

Conclusion Energy Multiband Result Introduction Energy consumption of e- ASA is increasing because

  • f number of users.

Energy usage of e-ASA and single are same for low number of users.

slide-20
SLIDE 20

Husnu S. Narman 24

Multi-shared-band with e-ASA

Improving throughput rate of multi-band up to 20% Up to 60% energy can be saved by using e-ASA.

Summary of Results

Conclusion Energy Multiband Result Introduction

slide-21
SLIDE 21

Husnu S. Narman 25

Conclusion

Conclusion Energy Multiband Result Introduction

slide-22
SLIDE 22

Thank You

http://students.ou.edu/N/Husnu.S.Narman-1 husnu@ou.edu

Husnu S. Narman 26