Energy%20Efficient%20Networking

1
Energy Efficient Networking

• Ramesh R. Rao
• University of California, San Diego
• - NeXtworking03 - Chania, Crete, Greece, June
  23-25,2003
• The First COST-IST(EU)-NSF(USA) Workshop on
  EXCHANGES TRENDS IN NETWORKING

2
Introduction

• Explosion of wireless devices and applications
• Eg. Cellular Networks, Ad Hoc Networks, Sensor
  Networks
• Power hungry applications and shrinking form
  factor
• Motivates need for energy aware designs

3
Three Themes

• Cross-layer optimization
• Resource ownership
• Batteries

4
Cross-layer Optimization

• Minimum energy/hops routing unsuitable
• Suppose R2 is chosen as the relay
• Lifetime Relay R2 will die prematurely
• Throughput More interference
• Need for joint routing and scheduling

5
The Problem is Hard

• Joint optimization is an NP-hard problem
• Arikan (1984)
• For small networks, a brute force approach works.
  
• Nuggehalli (2002)
• Need for scalable and distributed algorithms
  which can achieve near optimal performance

6
Lifetime Vs. Throughput Optimal Routes
Lifetime Optimal Routes Lifetime1331.6 Throughp
ut 1
Throughput Optimal Routes Lifetime839.2 Through
put1.33
7
Resource Ownership

• Ownership of resources determines protocol design
  paradigm
• Ad Hoc Network Distributed protocols that ensure
  system resources are used equitably and no user
  gets cheated

Bandwidth Energy
Cellular Networks System User
Ad Hoc Networks User User
Sensor Networks System System
8
Cooperation in Ad Hoc Networks

• Standard Assumption Nodes always relay messages
  for other nodes!
• Whats in it for me?
• No cooperation leads to zero throughput
• Complete cooperation leads to short active life.

S1
R
D
S2
S3
9
Cooperation in Ad Hoc Networks (Contd.)

• Two Questions
• How Much to relay?
• What strategy? (No cheating)
• Answer (Srinivasan etal Infocom 2003)
• Given complete information
• Each node calculates optimal level of cooperation
• Implements GTFT strategy (Nash Equilibrium)
• Challenge Learn operating point through
  experience with the system

10
Battery Management

• Maximize bits transmitted for a given battery
  wireless physical interface (both have serious
  impairments)
• Develop MAC protocols for both best battery life
  throughput (Tx when channel is best)
• Optimize Tx time, Rx time, Idle time, Sleep time
  (high power pulsed Tx time)
• Coordinate power consumption with battery state
  (decrease average power towards end of discharge)
• Operational scenario Wireless sensor network,
  solar powered by day, battery powered by night,
  sends data intermittently

11
Battery Test Set Up
12
Battery Test Results

• E-Tech Li-Ion Polymer Cell, 250mAh, 3.7Vdc
• Pulsed Discharge, 1 Second Period, 8C2A

13
Battery Tests, Summary

• Li-Ion polymer secondary cell, pulsed discharge,
  2x useful energy increase vs continuous discharge
  (charge recovery)
• Lithium coin primary cell, pulsed discharge, 8x
  useful energy increase vs continuous discharge
• Pulse constraints for charge recovery lt max on
  time, gtmin idle time, ltmax idle current, high
  pulse current levels possible, 8x to 80x
  manufacturers recommended current
• Improvement scenario Same weight, same power,
  much more energy or Same weight, much more power,
  same energy