Dissemination Ranges in Vehicular Networks by wuyunyi


									Data Dissemination in Vehicular Networks

                                 Vinod Kone

                               MAE Presentation

  Ben Zhao (co-chair), Heather Zheng (co-chair) & Elizabeth M. Belding
Motivation: Scenario I
Imagine traveling on a highway with traffic jam miles ahead…

                            • Data delivery latency is high
                            • High deployment cost for data collection

                   Can we do better?
    Motivation: Scenario II
   Imagine driving to a new city and you want to find the best parking lot

                                            Annual damage for a German city
                                            • € 20 million economic damage
                                            • € 3.5 million wastage on gasoline
                                            • 150k hours of waiting time

                                           Use a Parking Map!

                                          • No real-time information provided
                                          • e.g. current occupancy

                                  Use GPS POI
                                                Again, can we do better?

                                             We will revisit this !!!

So what kind of a system do we need?
   Desirable system properties
       Data collection and distribution in a local environment
       Low information delivery latency
       Cheap deployment and communication

   Probable solutions
       Cellular ? Service fees
       Satellite ? High latency
       Vehicular Networks ?

   What is a vehicular network?
       Vehicles are equipped with sensing, computing and wireless devices
       Vehicles talk to road-side infrastructure (V2I) and other vehicles (V2V)
       Has all the desirable properties
   Vehicular Networks and Applications

   Research overview

   Data dissemination approaches and tradeoffs

   Open problems

Vehicular Networks
   What does road-side infrastructure (Infostation)
       High bandwidth & Low cost device
       Coverage is less compared to a cellular base station

   Advantages of infrastructure support
       Low latency communication with vehicles
       Gateway to the Internet and extend connectivity
       Distributing time-critical data (e.g. accident notifications,
        traffic jam) near the affected area is efficient

Applications & Challenges
   Lots of potential applications
       Safety: Emergency brake light, Collision warning etc
       Comfort: Toll services, Parking space locator etc
       Commercial: Map updates, Video download, Services etc

   What makes vehicular networks challenging?
       Combination of V2V + V2I communications
       They face the ‘worst case’ scenarios in real world

Who all are working on vehicular networks?

  Automobile Industry              US DOT

                                    • Vehicle to Infrastructure test-bed, SFO
                                    • PATH, CarTel, DieselNet (USA)
                                    • FleetNet, NOW, CarTalk2000 (Europe)


                   http://roadtalk.wordpress.com                            8
Research on Vehicular Networks: The BIG Picture

    • DSRC                                            • Talk Focus
    • IEEE 802.11p

                        MAC &             Data
                     Physical Layer   Dissemination

                                       Security &
                      Models &

    • Traffic                                         •   Authentication
      Simulators                                      •   False Data
    • Real Traces                                     •   DOS Attacks
                                                      •   Privacy

MAC & Physical Layer
   Dedicated Short Range Communications (DSRC)
       Protocol for vehicles to talk to each other and infrastructure
       Operates in 75 MHz licensed band at 5.9 GHz (USA), 5.8GHz
        (Europe and Japan)

   Characteristics
       Based on 802.11a PHY and 802.11 MAC
       Supports high mobility of vehicles (120 mph)
       High data rate (27 Mbps), short range (1 km), multi-channel (7)

   Studies have shown that vehicle-to-infrastructure
    communication is feasible [Ott’04] [Bychkovsky’06]

   Vehicular Networks and Applications

   Research overview

   Data dissemination approaches and tradeoffs

   Open problems

Data Dissemination
Vehicular networks need to handle large amounts of
data (emergency messages, videos etc)

How do we efficiently disseminate this information?
Characteristics               Challenges
 High mobility                Maintaining routing tables is

 Dynamic topology              difficult
 Receivers are a priori
 Large scale                    Scalability
 High density

 Low penetration ratio          Dealing with partitions

Classification of Dissemination Approaches
   V2I / I2V dissemination
       Push based
       Pull based

   V2V dissemination
       Flooding
       Relaying

   How to deal with network partitions?
       Opportunistic forwarding

Push based dissemination

   Infostation pushes out the data to everyone
   Applications: Traffic alerts, Weather alerts

   Why is this useful?
       Good for popular data
       No cross traffic  Low contention

   Drawback
       Everyone might not be interested in the same data

Pull based dissemination

   Request – Response model
   Applications: Email, Webpage requests

   Why is this useful?
       For unpopular / user-specific data

   Drawback
       Lots of cross traffic  Contention, Interference, Collisions
Classification of Dissemination Approaches
   V2I / I2V dissemination
       Push based
       Pull based

   V2V dissemination
       Flooding
           Challenges
           Solutions & Drawbacks / Limitations
           Discovery of Parking Places problem
       Relaying

   How to deal with network partitions           ?
       Opportunistic forwarding
   Basic Idea
       Broadcast generated and received data to neighbors
       Usually everyone participates in dissemination

   Advantages
       “Good” for delay sensitive applications
       Suitable for sparse networks

   Key Challenges
       How to avoid broadcast storm problem?

Techniques to avoid the broadcast problem
   Simple forwarding
       Timer based [Linda’00]
       Hop limited [Nandan’06]

   Map based / Geographic forwarding
       Directed flooding [Sormani’06]
       Aggregation [Wischhof’04] [Nadeem’06] [Caliskan’06]

Drawbacks / Limitations of Flooding
   Flooding in general
       High message overhead  Not scalable

   Map based / Geographic
       Geographically closest doesn’t necessarily reflect the best
       Depend on a location based service
       Aggregation techniques tradeoff with accuracy

Decentralized Discovery of Parking Places
   Push + Map based + Flooding solution [Caliskan’06]

   Parking lots periodically broadcast occupancy and price information to
    nearby vehicles

   City map is divided into a quad-tree like structure

                                                                             Figure Source: Caliskan- VANETt’06
Decentralized discovery algorithm
   Information of a single parking lot is distributed only
    in proximity

   Aggregate information of a region is distributed over
    wide area

   Why this particular solution?
       Lots of vehicles are interested in the data  Push
       Fast transmission of the information  Flooding
       To avoid broadcast storm  Map based
Classification of Dissemination Approaches
   V2I / I2V dissemination
       Push based
       Pull based

   V2V dissemination
       Flooding
       Relaying
           2 Challenges
           Solutions & Drawbacks / Limitations

   How to deal with network partitions?
       Opportunistic forwarding
   Basic Idea
       Instead of flooding the network, select a relay (next hop)
       Relay node forwards the data to next hop and so on

   Advantages
       Reduced contention  Scalable for dense networks

   Key Challenges
       How to select the relay neighbors?
       How to ensure reliability?

How to select a relay neighbor?
   Simple forwarding
       Select the node farthest from source [Korkmaz’04]
        [Zhao’07] [Our work]

   Map based / Geographic forwarding
       Closest to the destination [Kikaiakos’05]
       Abstract topology into a weighted directed graph [Zhao’06]

   Drawback / Limitations
       Locally best next hop may not be globally best !
How to ensure reliability?
   Use RTS/CTS & ACK [Korkamaz’01] [Zhao’07]

   Use indirect acknowledgments [Benslimane’04]
    [our work]

   Drawbacks / Limitations
       RTS/CTS incurs lot of overhead
       Interference affects indirect acknowledgments

Classification of Dissemination Approaches
   V2I / I2V dissemination
       Push based
       Pull based

   V2V dissemination
       Flooding
       Relaying

   How to deal with network partitions?
       Opportunistic forwarding

Opportunistic Forwarding
   Problem with partitioned networks
       Next hop is not always present

   Opportunistic Forwarding
       Basic Idea: Store and Forward
       Challenge: What is the right re-broadcast interval?

   Solutions
       Broadcast repeatedly [Linda’00b][Uichin’06][Wischhof’04]
       Cache at infostations [Lochert’07a]

Opportunistic: Drawbacks / Limitations

   It is difficult to select the correct re-broadcast interval
       Too soon  high overhead
       Too late  doesn’t deal with partitions effectively

   Maintaining a neighbor list induces high overhead
    and contention

     Dissemination Approaches: The BIG Picture
Map based/

                                            * PULL
Take Away
V2I/I2V Dissemination   Pros                        Cons

Push                    Suitable for popular data   Not suitable for un-popular
Pull                    Suitable for un-popular/    Cross traffic incurs heavy
                        user-specific data          interference, collisions

V2V Dissemination       Pros                        Cons
Flooding                Can reliably & quickly      Not scalable for dense
                        distribute data             networks
Relaying                Works well even in dense    Selecting best next hop &
                        networks                    reliability is difficult

Dissemination in        Pros                        Cons
Partitioned networks
Opportunistic           Suitable for network        Difficult to estimate
                        partitions                  re-broadcast interval
                                                    High overhead in dense
                                                    networks                     30
   Vehicular Networks and Applications

   Research overview

   Data dissemination approaches and tradeoffs

   Open problems

Some interesting open problems
   Not much literature on V2I / I2V communication
       How to deal with cross-traffic in the pull scheme
       Scheduling transmissions?
       How to combine push and pull ? What is hybrid ?

   Mobility traces for evaluation of dissemination
       Real traces (e.g. NGSIM) are expensive to collect
       Not enough data points for simulation
       Need to extrapolate

Some interesting open problems                              (contd…)

   Imagine a service provider wants to install
       What is the minimum infostation density required
       Impact of application parameters (size, lifetime)

   Analytical models
       Understand the bounds on performance
       Modeling network partitions  Better opportunistic

Some interesting open problems                             (contd…)

   Real experiments
       Equip vehicles with wireless devices and observe
        dissemination performance
       Can obtain real movement traces
       Designing and testing sample applications

   Real experiments might invalidate the design!
       Re-design the schemes based on the real observations
       Repeat!

Future Work
   Hybrid dissemination in vehicular networks

   Developing accurate analytical dissemination models

   Real experiments

Thank You for Listening

   [Nadeem’06] Comparative study of data dissemination models for vanets, Mobiquitous.
   [Wu’04a] MDDV: A Mobility-Centric Data Dissemination Algorithm for Vehicular
    Networks, VANET.
   [Korkamaz’04] Urban multi-hop broadcast protocol for inter-vehicle communication
    systems, VANET.
   [Sun’00] GPS-Based Message Broadcasting for Inter-Vehicle Communication, ICCPP.
   [Zong’01] Ad Hoc Relay Wireless Networks over Moving Vehicles on Highways,
   [Wu’04b] Analytical Models for Information Propagation in Vehicle-to-Vehicle
    Networks, VTC.
   [Linda’00a] Disseminating Messages among Highly Mobile Hosts based on Inter-
    Vehicle Communication, IV.
   [Sormani’06] Towards Lightweight Information Dissemination in Inter-Vehicular
    Networks, VANET
   [Zhao’06] VADD-Vehicle-Assisted Data Delivery in Vehicular Ad Hoc Networks,
   [Caliskan’06] Decentralized Discovery of Free Parking Spaces, VANET
   [Basu’04] Wireless Ad Hoc Discovery of Parking Meters, WAMES.
   [Zhao’07] Data Pouring and Buffering on The Road: A New Data Dissemination
    Paradigm for Vehicular Ad Hoc Networks, Transactions on VT

   [Lochert’07a] The Feasibility of Information Dissemination in Vehicular Ad-Hoc
    Networks, WON
   [Wischhof’04] Information Dissemination in Self-Organizing Inter-vehicle Networks,
    Trans on ITS
   [Uichin’06] FleaNet: A Virtual Market Place on Vehicular Networks, MobiQuitos
   [Kikaiakos’05] VITP: An information transfer protocol for vehicular computing,
   [Bai’06] Towards Characterizing and Classifying Communication-based Automotive
    Applications from a Wireless Networking Perspective, Research Report, GM
   [Nandan’06] Modeling Epidemic Query Dissemination in AdTorrent Network, CCNC
   [Linda’00b] Role-Based Multicast in Highly Mobile but Sparsely Connected Ad Hoc
    Networks, MobiHoc
   [Lochert’07b] Probabilistic Aggregation for Data Dissemination in VANETs, VANET
   [Luo’04] A Survey of Inter-Vehicle Communication, Technical Report
   [Varghese’06] Survey of Routing Protocols for Inter-Vehicle Communications,
   [Bychkovsky’06] A Measurement Study of Vehicular Internet Access Using In Situ Wi-
    Fi Networks, MobiCom
   [Choo’06] Performance Study of Robust Data Transfer Protocol for VANETs, LNCS.
   [Bensilmane’04] Optimized dissemination of alarm messages in vehicular ad-hoc
    networks, HSNMC
   [Bala’07] Web Search From a Bus, CHANTS.
   [Burgess’06] MaxProp: Routing for Vehicle-Based Disruption-Tolerant Networks,
   [Ott’04] Drive-thru Internet: IEEE 802.11b for “Automobile” Users, INFOCOM.
   [Hartenstein’01] Position-Aware Ad Hoc Wireless Networks for Inter-Vehicle
    Communications, MobiHoc
   [Namboodiri’04] A study on the feasibility of mobile gateways for vehicular ad-hoc
   [Shahram’04] PAVAN: A Policy Framework for Content Availability in Vehicular Ad-
    hoc Networks, VANET
   [Raya’07] Securing Vehicular Networks, INFOCOM
   [Harsch’06] Secure Position Based Routing for VANETs, VTC

   Traffic View [Nadeem’06]
       Formal models for data dissemination
       Bi-directional mobility considered
       Aggregation based flooding
       Not scalable for dense traffic
       Flooding is in general good for delay sensitive apps
       Targeted App: Traffic monitoring

   MDDV[Wu’04a]
       Opportunistic + trajectory + geographic forwarding
       Assumes vehicles have road map and know src, dest region
       Traffic flow information is fed to vehicles to abstract the road map and
        make forwarding decisions
       Group of vehicles near the message head can forward the data
       Forwarding phase to reach the destination region and then propagation
        phase to reach all the receivers in the region
   Adhoc Relay [Zong’01]
       Opportunistic (pessimistic) forwarding based on store and forward approach
       Good for networks with low density
       Delay-sensitive applications cannot work with this
       Motion of vehicles significantly affect delivery latency
   Analytical Models [Wu’04b]
       Model an idealistic propagation scheme
       Consider partitioning of vehicles for information propagation
       Forward (intra-partition) and catchup(inter-partition) processes
       Models are for sparse (ignores in-partition propagation) and dense (traffic
        between cycles) networks
       Doesn’t model real networks
   VITP [Kikaiakos’05]
       Geographical routing to forward the query to the query region
       Nodes maintain a neighbor list
       Once query region is reached, nodes do flooding
       Reply is sent back to the source via flooding
   Urban-Multihop [Korkamaz’01]
       Segment the road in the dissemination direction iteratively
       Select the node in the furthest segment as relay
       RTS/CTS like mechanism at MAC layer
       Distance from source decides black burst time
       Repeaters are used at intersections to propagate to different directions
   Dissemination Messages [Linda’00a]
       Flooding based solution
       Nodes wait a time proportional to the distance from the source before
   Role based Multicast [Linda’00b]
       [Linda’00a] + retransmissions based on change in the neighbor set
   Lightweight Dissemination [Sormani’06]
       Dissemination is based on propagation function
       Propagation function encodes destination region and trajectory
       Propose several flooding schemes (basic, probabilistic, function driven)
       Requires a map to create and evaluate the propagation function
   VADD [Zhao’06]
       Pull based routing model to query a static location
       Map based information (trajectory, traffic) is used to select the next hop with least delay to the
       Models roads and intersections as graphs with estimated delays as weights
       Store and forward approach to tackle sparse networks
       Targeted App: To query a static information center
   DP, DP-IB [Zhao’07]
       Propose a data pouring and buffering dissemination scheme
       Nodes maintain neighbor list and select farthest node as relay
       RTS/CTS, Indirect Acks are used for reliability
       Ibers (Infostations) are deployed at intersections to rebroadcast data on the cross roads
       Analytical models developed for dissemination capacity and broadcast interval
   Feasibility [Lochert’07]
       Shows the feasbility of information dissemination w.r.t. penetration ratio in city
       Analytical model to show that connectivity decreases with length
       Propose installing SSUs (InfoStations), networked and stand-alone to improve dissemination by
        re-broadcasting the information
       Vehicles periodically broadcast information to neighbors (Locomotion + Wireless propagation)

   SODAD/SOTIS [Wischhof’04]
       Data dissemination is achieved by abstracting the map into segments
        and aggregating information
       Analytical models (coverage processes) to show low penetration ratio
        leads to low multi-hop range
       Recurrent broadcasts to tackle with network partitions
       Adaptive broadcast interval based on provocation/mollification events
        to suit traffic conditions
       Targeted App: Vehicles sensing data for traffic info system
   FleaNet [Uichin’06]
       Proposed an architecture for buy/sell queries dissemination
       Dissemination is basically by contacts…vehicles that receive queries
        store it in their db and see if there is a local match
       Source broadcasts queries periodically to its neighbors (opportunistic)
       LER routing is used to send notifications from buyers sellers
Some stats
   Number of telemetric subscribers will reach
    >15 million by 2009

   Smart traffic lights can reduce waiting time by 28%
    during rush hours

Mobility Models & Simulators
   How to evaluate vehicular network protocols?
       Synthetic mobility models: highly unrealistic
       Trace-driven

   Traces from microscopic traffic simulators
       close to reality but not real

   Real Traces (Source: NGSIM)
       very expensive to collect data
       not enough data points

   How can we solve this problem?
       We have to extrapolate the real data by some “modeling”
       Equip vehicles with sensors
Security and Privacy: Why is this important?

                                                                          Network Disruption
Bogus Traffic Information


                            Figures borrowed from Hubaux Sevecom
Secure solutions for VANETs
                        PKI, Shared Keys

[Raya’07] [Harsch’06]

                                    Data Correlation

For more info:
• http://ivc.epfl.ch    Switch channels
• http://sevecom.org


To top