Context-Driven Disruption Tolerant Networking for Vehicle Applications

Transcription

1 Cntext-Driven Disruptin Tlerant Netwrking fr Vehicle Applicatins 1 Wai Chen Directr and Chief Scientist Applied Research, Telcrdia Technlgies Cntact: t waichen@ieee.rg

2 Cntext-Driven Disruptin Tlerant Netwrking fr Vehicle Applicatins Wai Chen 1, Jasmine Chennikara-Varghese 1, Ratul Guha 1, Jhn Lee 1, Rykichi Onishi 2, Rama Vuyyuru 2, Junichir Fukuyama 2 1 Telcrdia Technlgies 2 Tyta InfTechnlgy Center March 3-4,

3 Backgrund and Mtivatin fr DTN 3 Disseminatin f infrmatin amng vehicles and RSUs with frequent cnnectivity disruptins Current appraches adapted frm P2P cntent delivery netwrks Cnnectin variants Vehicle-based Peer-t-Peer RSU-assisted Peer-t-Peer Backbne-assisted Peer-t-Peer RSU Appraches Fragmentatin and frwarding Carry and frward peratin Practive cnfiguratin Challenges Lng duratin infrmatin dwnlad Dynamic Vehicle Psitin Frequent Discnnectin f Links RSU

4 Sme Related Wrk Gegraphical Opprtunistic Ruting fr Vehicular Netwrks (GeOps) Apprach Explits gegraphic infrmatin frm vehicle navigatin system t deliver message t particular lcatin. Given the destinatin lcatin, the sending vehicle (carrier) cmpute the nearest pint n its rute t the destinatin. Based n nearest pint it will calculate the minimum estimated delivery time t the destinatin by using navigatin system. 4 Each vehicle will cmpute the minimum estimated delivery time based n its rute. And vehicle C will have minimum estimated delivery time and chse t be carrier f message. Vehicle e C will chse anther carrier t transmit t the message t the destinatin University Cllege, Lndn

5 Sme Related Wrk 5 DTN Prtcl Destinatin Applicatins GeOpps Chse frwarding that will take message clse t destinatin based n navigatin system. Fixed single pint Nt specified GeDTN+Nav Nav Use Ge-graphic ruting and switch t DTN mde when netwrk partitins are detected Determine delivery quality metric based n navigatin system t chse best neighbr Fixed single pint Same as abve VADD Use predictable vehicle mbility t chse best carry-and-frward neighbr. Fixed single pint Nt specified (Packet ruting) Cntext-driven DTN Use cntext infrmatin t ptimize and disseminate cntent Mbile multiple pints Driving assistance Cntext-driven DTN (C-DTN): Netwrking prtcl peratins driven by cntext In cntrast t a generic mbile backbne t achieve end-t-end netwrking via multi-hp disseminatin

6 Design Apprach f Cntext-driven DTN (C-DTN) Cntext Identificatin 6 Fcus n individual vehicle applicatin sessins and requirements Leverage cntext t develp technlgy fr disruptin tlerant netwrk. Cntext cnsists f Tempral scpe Cntent expiratin, (Packet expiratin) Spatial scpe Cntent psitin, directin, trajectry, area (regin f interest, psting area) Driving Attributes Vehicle infrmatin including trajectry (driving plan) Used fr estimatin f future lcatins Cmmunicatin pattern End-t-end, end-t-multiple ends, end-t-regin, etc. Etc.

7 Infrmatin Flw in System Cmpnents 7 Cntext-driven DTN Cntent RSU Cntent querying RSU Cntext generatr Cntent querying Infrastructure Cntext handling Disruptin Tlerant Netwrk Cntent generatin Cntent replying Cntent psting

8 Lgical Stack Cmpnents f C-DTN 8 Applicatins Cntext Handling Sub-layer Cntext Generatin Cntext Infrmatin Mdule Mdule Cntent Optimizatin Sub-layer Detectin & Cnditin Mdule DTN Frwarding Mdule DTN Sub-layer Packet IN Packet OUT

9 Evaluatin Scenaris Overview Scenari # 1 Cntent Surce: static single nde Cntent Recipients: static intersectins Effect f variatin in vehicle density and mbility n perfrmance Scenari # 2 Cntent Surce: multiple intersectins Cntent Recipient: mbile vehicles Effect f variatin in vehicle density and destinatin mbility n perfrmance 9 D D S S 513 D 514 D S D Sender Destinatin lcatins S S Carrying vehicle Receiving Vehicle 501 S Applicatin Parameter Cumulative Packet Rate Radi Range Duratin f simulatin Speed distributin Value 1300 B 1 packet/sec 100m 600 secnds Unifrm [24-96] Kmph

10 Scenari Evaluatin f C-DTN Packet Delivery Rati and Average Delay Scenari 1 PDR fr #519 Scenari 2 PDR frm Surce Delivery Rati fr 1 way; 1 lane 519 fr 1 way;2 lane 519 fr 2 way; 2 lane Delivery Rati fr 1 way; 1 lane 519 fr 1 way; 2 lane 519 fr 2 way; 2 lane Vehicle Density (vehicles per mile) Cmparisn fr PDR at #519 (farthest frm surce) Scenari 1 Delay fr # Vehicle Density (vehicles per mile) Cmparisn f PDR at mbile vehicle 1 frm surce #519 Scenari 2 Delay fr Surce fr 1 way; 1 lane 519 fr 1 way; 2 lane 519 fr 2 way; 2 lane Secnds Secnds fr 1 way; 1 lane 519 fr 1 way; 2 lane 519 fr 2 way; 2 lane Vehicle Density (Vehicles per mile) Vehicle Density (vehicles per mile) Geneva, Cmparisn 3-4 fr March delay 2010 at #519 (farthest frm surce) Cmparisn f delay at mbile vehicle 1 frm surce #

11 Scenari Evaluatin f C-DTN Freshness Vehicle density = 5 vehicles per mile 11 2 way; 2 lane 2 way; 2 lane Sent by surce Received at #519 DTNMP #519 (farthest frm surce) receives frm cntent surce Distinct instances f cntent receptin Mbile destinatin receives frm farthest cntent surce Mbile destinatin receives cntent mre frequently due t crrelated mbility

12 Summary 12 Overviewed general DTN appraches Prpsed a cntext-driven DTN (C-DTN) that supprts netwrking in cnsideratin f applicatin cntext needs Evaluatins t validate C-DTN disseminatin perfrmance with realistic mbility Diversity in mbility pattern generally imprves perfrmance Under crrelated mbility mdel, DTN disseminatin perfrmance depends n a number f factrs

13 Backup Slides

14 Sme Related Wrk GeDTN+Nav: Hybrid Gegraphic and DTN Ruting with Navigatin Assistance Apprach Start frwarding in greedy frwarding (shrtest path) mde and switch t DTN when netwrk is discnnected Prpsed scre functin t detect the netwrk discnnectin. Cmpute delivery quality metric fr each neighbr and chse the best quality nde t frward the packet. Ex. Cmpute d1, d2 and d3 fr 3 neighbrs and chse best (d1) nde. (belw) 14 d2 Prpsed Vehicle Navigatin Interface (VNI) fr these vehicles. d3 d1 Univ. Of Califrnia, Ls Angeles, 2008

15 Sme Related Wrk Vehicle-Assisted Data Delivery (VADD) Prtcl 15 Apprach Estimates packet delivery delay n each radway. Select frwarding directin using a pririty list based n prbabilistically shrtest delay path cmputed in a centralized manner. Surce and destinatin assumed t be in a cnnected graph Gegraphically shrtest path (traditinal) 1 Penn State University, PA, Fast speed wireless cmmunicatin path (VADD)