17
CSLI 5350G - Pervasive and Mobile Computing Week 5 - Paper Presentation “An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks” Name: Dinesh Bilimoria Date: Oct 16 th , 2013

Name : Dinesh Bilimoria Date: Oct 16 th , 2013

  • Upload
    landen

  • View
    33

  • Download
    0

Embed Size (px)

DESCRIPTION

CSLI 5350G - Pervasive and Mobile Computing Week 5 - Paper Presentation “An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks”. Name : Dinesh Bilimoria Date: Oct 16 th , 2013. Research Paper. Bibliography: - PowerPoint PPT Presentation

Citation preview

Page 1: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

CSLI 5350G - Pervasive and Mobile ComputingWeek 5 - Paper Presentation“An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks”

Name: Dinesh BilimoriaDate: Oct 16th, 2013

Page 2: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

2

Research Paper

Bibliography:

Villas, L.A., Ramos, H. S., Boukerche, A., Guidoni, D., Araujo, B,R., and Loureiro, A.A.F.(2012). "An efficient and robust data

dissemination protocol for vehicular ad hoc networks." In Proceedings of the 9th ACM symposium on Performance evaluation of wireless ad hoc, sensor, and ubiquitous networks (PE-WASUN '12). ACM, New York, NY, USA, 39-46.

Page 3: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

3

Background

Urban Multihop Broadcast (UMB)

Vehicle

Repeater

Page 4: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

4

Background

Distributed Vehicular Broadcast (DV-CAST)

Well Connected Neighborhood(one neighbor is in message forwarding direction -> broadcast suppression)

Sparsely Connected Neighborhood(last node in cluster, and one neighbor in opposite direction-> store and forward)

Page 5: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

5

Background

Distributed Robust Geocast (DRG)Two-dimensional ZOR - the relay nodes should have a wider angular distance to cover new regions of ZOR

Non-realistic Model Realistic Model

Page 6: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

6

Background

Mobicast

Page 7: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

7

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Objective The goal to provide solution to broadcast storm problem, network partition

and temporal network fragmentation.

Proposal GEographical Data Dissemination for Alert Information (GEDDAI) Eliminate the broadcast storm Maximizes the capability of performing data dissemination across ZoR Low overhead, short delays and high coverage

Page 8: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

8

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Contributions The solution performs data dissemination in a reactive way, i.e., it does not

require the building and maintenance of a table of neighbors in order to disseminate data across a ZoR

Proposed a sweet spot (ie ZoP) and use it to efficiently disseminate the data across a zone of relevance achieving low overhead, short delays and high coverage

Assumptions Perform data dissemination with ZoR Vehicles do not require table of neighbors Do not need to use ZoF

Page 9: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Sweet Spot (Zone of Preference) An area where the vehicles are inside it to perform data dissemination

Page 10: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

10

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Flow Chart Diagram

Page 11: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

11

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Data Dissemination

When an event occurs source vehicle starts to disseminate data about the eventSource vehicle floods the ZoR with AI message and forwards to neighborsOne neighbor node is chosen to continue data disseminationIf no vehicles in Sweet Spot, the vehicle farther from the transmitter will be chosenThe process is performed until information is disseminated throughout

Page 12: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

12

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Evidence Use of metrics - overhead, collisions, coverage and delay

Evaluation of Prototype OMNeT++ NS v4.1 Inetmanet Framework SUMO

Page 13: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

13

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Results

Page 14: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

14

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Results

Page 15: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

15

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Shoulder of Giants Distributed Vehicular Broadcast (DV-CAST) Distributed Robust Geocast (DRG) Adaptive Information Dissemination (AID) Distance Based Relay Selection (DBRS)

Impact Cited by 3 since 2012

Page 16: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

16

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Open problems Analytical and modeling metrics to determine if proposing message alert will

be practical in sweet spot (ZoP) Flow chart should indicate retry method before message is cancelled Impact of sweet spot (ZoP) message dissemination if communication radius

is more then 200 meters

Page 17: Name : Dinesh Bilimoria Date:    Oct 16 th ,  2013

17

An Efficient and Robust Data Dissemination Protocol for Vehicular Ad Hoc Networks

Discussion points1. What is a Sweet Spot? An area where the vehicles are inside it to perform data dissemination

2. What method is mentioned in the paper to eliminate broadcast storm? GEographical Data Dissemination for Alert Information (GEDDAI)

3. What is the impact to the results on Page 12 and Page 13 if the dissemination rate is increased from 1 sec to 2 secs. Message will take almost twice the time to disseminate since overhead

increases.