A Multicast Protocol for Mobile Ad Hoc A Multicast Protocol for - - PowerPoint PPT Presentation

a multicast protocol for mobile ad hoc a multicast
SMART_READER_LITE
LIVE PREVIEW

A Multicast Protocol for Mobile Ad Hoc A Multicast Protocol for - - PowerPoint PPT Presentation

4 th Workshop on Applications and Services in Wireless Networks A Multicast Protocol for Mobile Ad Hoc A Multicast Protocol for Mobile Ad Hoc Networks Using Location Information Networks Using Location Information Habib Ammari and Hesham


slide-1
SLIDE 1

Computer Science and Engineering

4th Workshop on Applications and Services in Wireless Networks

A Multicast Protocol for Mobile Ad Hoc A Multicast Protocol for Mobile Ad Hoc Networks Using Location Information Networks Using Location Information

Habib Ammari and Hesham El-Rewini

{hammari,rewini}@engr.smu.edu

Boston University, Boston, Massachusetts, USA August 8-11, 2004

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

slide-2
SLIDE 2

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Introduction and Motivations Background Multicast Protocol Design Illustrative Example Summary and Perspectives

Outline Outline

slide-3
SLIDE 3

Computer Science and Engineering

Definition: A mobile ad hoc network (MANET) is a collection of mobile nodes without any infrastructure Mobile node behavior

  • Mobile nodes act as hosts (running applications) and

routers (forwarding for others) MANET architectural properties

  • Autonomous nodes
  • Distributed operation
  • Multihop routing
  • Dynamic topology
  • Limited capabilities

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Introduction and Motivations

slide-4
SLIDE 4

Computer Science and Engineering

MANET routing protocols

  • Proactive vs. reactive
  • Unicast vs. multicast
  • One-to-one: unicasting •

One-to-many: multicasting Ad hoc networking applications

  • Establishing infrastructured networks is impossible
  • r not cost effective
  • Temporary networks for urgent situations such as

battlefields, earthquake, conferencing, etc.

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Introduction and Motivations (cont’d)

slide-5
SLIDE 5

Computer Science and Engineering

Why multicast routing protocols?

  • Same message sent to a group of mobile nodes
  • Group communication in military applications

Our objective

  • Develop a multicast protocol for MANETs
  • Minimize routing overhead

Tools

  • Location information (GPS-enabled mobile nodes)
  • Voronoi diagrams structural properties

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Introduction and Motivations (cont’d)

slide-6
SLIDE 6

Computer Science and Engineering

Voronoi diagram

  • Geometrical construct defined by a discrete set of

sites (points) S = { s1,s2,s3,…,sn } in the plane

  • Nearest-neighbor rule: each point is assigned with

the closest region of the plane to it

  • B(si ,sj ) = { p∈ℜ 2 | δ(si ,p) = δ(sj ,p) : si ,sj∈S}: bisector
  • f si and sj in S, where δ: Euclidean distance function

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Background

slide-7
SLIDE 7

Computer Science and Engineering

  • HP(si ,sj ) = {p∈ℜ 2 | δ(si ,p) < δ(sj ,p) : si ,sj∈S}
  • HP(sj ,si ) = {p∈ℜ 2 | δ(sj ,p) < δ(si ,p) : si ,sj∈S}
  • VR(si ,S) = ∩ HP(si ,sj ): Voronoi region of si
  • Boundary of a Voronoi region: Voronoi edges
  • Endpoints of a Voronoi edge: Voronoi vertices

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Background (cont’d)

slide-8
SLIDE 8

Computer Science and Engineering

  • The boundary of a region has at most n-1

Voronoi edges

  • Voronoi regions constitute a polygonal partition
  • f the plane: Voronoi diagram V(S)
  • V(S) = ∪ VR(si ,S): Voronoi diagram of S

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Background (cont’d)

slide-9
SLIDE 9

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Background (cont’d)

Illustrative Examples

slide-10
SLIDE 10

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Background (cont’d)

Voronoi Diagram Construction Algorithms

  • Straightforward approach – construct one region

at a time as the intersection of n-1 half-planes ⇒ O(n2) time for one region ⇒ O(n3) time algorithm

  • Divide-and-conquer algorithm ⇒ O(n log n)
  • Shamos and Hoey ⇒ O(n log n) time algorithm
slide-11
SLIDE 11

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Multicast Protocol Design

MANET modeling

  • MANETs can be modeled using Voronoi diagram

Neighboring node set

  • NN(si ): neighboring node set of si is the set of

MANET nodes within the transmission range of si Assumptions

  • GPS: location information to MANET nodes
slide-12
SLIDE 12

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Multicast Protocol Design (cont’d)

  • MANET node si broadcasts its location

information when joining MANET

  • When changing its location, a MANET node

might have to broadcast its location information ⇒ decision based on its new location and current distances to its neighboring nodes

  • Neighboring nodes reply back with their location

information with time-to-live = 1

slide-13
SLIDE 13

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Multicast domain

  • Planar region
  • ζ = (MD,ρMD), where

MD = (xMD,yMD)

  • Membership to the

multicast group δ(si,MD) ≤ ρMD

  • {s5,s8,s10}: multicast

group wrt to s1

Multicast Protocol Design (cont’d)

slide-14
SLIDE 14

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Authorized Forwarders

  • MANET node sa∈NN(ss ) is an authorized

forwarder of a multicast packet broadcast by ss if sa’s Voronoi region share at least one Voronoi edge with that of MD in ss’s localized Voronoi diagram VGss wrt to NN(ss )∪ {ss }∪ {MD }

  • sa constructs its localized Voronoi diagram wrt to

NN(sa ) ∪ {sa } ∪ {MD} \ (AF(scs ) ∪ {scs })

Multicast Protocol Design (cont’d)

slide-15
SLIDE 15

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Algorithm

Multicast Protocol Design (cont’d)

slide-16
SLIDE 16

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

  • Assume s1 wants to send a

multicast packet to ζ = (MD,ρMD)

  • NN(s1 ) = {s2,s3,s4,s6,s7,s9 }

NN(s2 ) = {s1,s3,s5,s6,s8 } NN(s6 ) = {s1,s2,s3,s5,s8,s9,s10 } NN(s9 ) = {s1,s3,s4,s6,s7,s8,s11 }

Illustrative Example VGs1

slide-17
SLIDE 17

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Illustrative Example (cont’d) VGs6 VGs9

slide-18
SLIDE 18

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

Illustrative Example (cont’d)

Blank space between source and multicast domain

  • Location-based multicast

(geocasting) provided by Y. Ko and N. Vaidya: efficient geographical multicast protocol for MANETs

  • Their algorithm is based on

location and distances between mobile nodes

  • It fails in case of blank

space between source and multicast domain

slide-19
SLIDE 19

Computer Science and Engineering

  • Multicast protocol for mobile ad hoc networks
  • Location information and structural properties of

Voronoi diagrams (authorized forwarders to reduce the routing overhead)

  • Mathematical analysis of the proposed protocol
  • Simulation of the proposed protocol using different

mobility models (RWP, RPGM)

  • Using this protocol in the integration of the MANETs

and the global Internet (providing mobile services)

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab

Summary and Perspectives

Computer Science and Engineering Computer Science and Engineering

slide-20
SLIDE 20

Computer Science and Engineering

  • F. Aurenhammer, “Voronoi Diagrams – A Survey of a

Fundamental Geometric Data Structure,” ACM Computing Survey, Vol. 23, No. 3, September 1991.

  • Y. Ko and N. Vaidya, “Geocasting in Mobile Ad hoc Networks:

Location-Based Multicast Algorithms,” Second IEEE Workshop

  • n Mobile Computer Systems and Applications, New Orleans,

Louisiana, USA, February 25-26, 1999

  • M. Shamos and D. Hoey, “Closest-Point Problems,” Proceedings of

the 16th Annual IEEE Symposium on Foundations of Computer Science, The University of California, Berkeley, USA, 13-15 October 1975.

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab

Useful References

Computer Science and Engineering Computer Science and Engineering

slide-21
SLIDE 21

Computer Science and Engineering

Mobile Computing and Parallel Processing Lab Mobile Computing and Parallel Processing Lab Computer Science and Engineering Computer Science and Engineering

A Multicast Protocol for Mobile Ad Hoc A Multicast Protocol for Mobile Ad Hoc Networks Using Location Information* Networks Using Location Information*

Habib Ammari and Hesham El-Rewini

{hammari,rewini}@engr.smu.edu

* Work supported by the Department of Defense (DoD)

Thank you! Questions?