Communication Networks

1
Communication Networks

• A Second Course

Jean Walrand Department of EECS University of
California at Berkeley
2
Administrative Details

• 3 Units
• Course Contents
• Review major recent networking research ideas
• Provide background for these papers
• Get a sense of where field is heading
• Grade based on
• Class participation Writing lecture notes
• Lecture ? assigned scribe writes notes ? other
  students send comments on notes ? revised notes
• Term project and its presentation
• Select topic by week 6 outline by week 9 ready
  to present by week 12 report by week 15.

3
Contents

• Introduction
• Review of TCP/IP
• Moving Forward
• Wireless Networks
• Transport
• Incentives

4
Introduction

• Explain network models and analysis
• Focus on three active areas of research
• Wireless Wi-Fi, Wi-Max, MANETs, Sensors
• QoS, Transport, Routing, Cross-Layer
• Incentives Differentiation, Revenue Sharing
• Along the way, we cover some useful tools
• Markov models
• Stability of nonlinear systems
• Duality in convex programming
• Game theory

5
Introduction

• Major developments in Internet and networks
• Economic value (B2B, B2C)
• New applications (voice, video, TV)
• New structures (P2P)
• Machine-to-machine communication
• Mobility support
• Security concerns
• Internet protocols over-stretched
• Protocols must adapt or be changed

6
Introduction

• Key Ideas

7
Introduction

• Major research topics

8
Introduction

• Architecture
• Service Model Best effort or richer?
• Compatible or Greenfield
• New layer structure for wireless, cross-layer?
• Switching and transport modes

9
Introduction

• Architecture Service

Specification soft or precise? Choice and select?
10
Introduction

• Architecture Layers

11
Introduction

• Architecture Mode

12
Introduction

• Middleware
• Economics Differentiation, Revenue Sharing
• QoS Specified or end-to-end choice
• Mobility Indirection Name ? Address
• Security Insurance?

13
Introduction

• Middleware Differentiation

14
Introduction

• Middleware QoS
• Specified Premium less than 150ms
• Choice premium better than basic

15
Introduction

• Middleware Mobility
• Indirection (similar to cell phone)

16
Introduction

• Middleware Security
• Insurance?To connect, you need an
  insuranceInsurance company checks your
  securityIncentive Premium goes down if secure

17
Introduction

• Wireless
• Sensor and Ad Hoc
• Application-driven or generic?
• Network capacity
• Scheduling graph coloring
• Power control strategic issues
• Routing interference short or long hops
  parallel paths
• Transport How to include graph constraints?
• Network-processing
• Network coding
• Wi-Fi
• VoIPData Capacity
• WiMax vs. Wi-Fi

18
Introduction

• Wireless Application-Driven or Generic?
• Basic observation No universally best protocols
• Using features of application yields better
  solutions
• Example 1 If central node can reach all sensors,
  it can provide synchronization, polling, etc.
• Example 2 Power constraints may not exist in
  some systems and be essential in others
• Keep in mind technology trends

19
Introduction

• Wireless Network Capacity
• Random network or given network?
• O(.)-results or precise results?
• Optimal or given set of protocols?
• Delay-constrained or not?

20
Introduction

• Wireless Scheduling
• Graph coloring
• Key issue correct model of constraints
• Difficulty InterferenceUncertainties
• Question Complexity (time, communication)

21
Introduction

• Wireless Power Control
• Strategic In CDMA, increasing a improves rate
  for node but increases interference for others

22
Introduction

• Wireless Routing
• Principle of optimality fails.
• Short or long hops?

23
Introduction

• Wireless Transport
• Constraints?Assume CDMA or TDMA?

24
Introduction

• Wireless Network Processing
• Complexity

25
Introduction

• Wireless Network Coding
• Worthwhile?

26
Introduction

• Wired
• Overlay pros and cons
• P2P incentives
• Transport Control problem, strategic aspects
• Service differentiation benefits
• Routing BGP, QoS, Strategic aspects