Proposed ad hoc Routing Approaches• Conventional wired-type schemes (global

routing, proactive):– Distance Vector; Link State

• Proactive ad hoc routing:– OLSR, TBRPF

• On- Demand, reactive routing:• DSR (Source routing), MSR • AODV (Backward learning)

• Scalable routing :– Hierarchical routing: HSR, Fisheye– OLSR + Fisheye– LANMAR (for teams/swarms)

• Geo-routing: • GPSR, GeRaF, etc• Motion assisted routing• Direction Forwarding

Georouting - Key Idea

• Each node knows its geo-coordinates (eg, from GPS or Galileo)

• Source knows destination geo-coordinates; it stamps them in the packet

• Geo-forwarding: at each hop, the packet is forwarded to the neighbor closest to destination

• Options:– Each node keeps track of neighbor

coordinates– Nodes know nothing about neighbor

coordinates

Greedy Perimeter Stateless Routing for Wireless Networks (GPSR)

– key elements• Greedy forwarding

– Each nodes knows own coordinates– Source knows coordinates of destination– Greedy choice – “select” the most forward

node

Finding the most forward neighbor

• Beaconing: periodically each node broadcasts to neighbors own {MAC ID, IP ID, geo coordinates}

• Each data packet piggybacks sender coordinates

• Alternatively (for low energy, low duty cycle ops) the sender solicits “beacons” with “neighbor request” packets

Greedy Perimeter Forwarding

D is the destination; x is the node where the packet enters perimeter mode; forwarding hops are solid arrows;

Got stuck? Perimeter forwarding

> Greedy forwarding failure. x is a local maximum in its geographic proximity to D; w and y are farther from D.> Node x’s void with respect to destination D

GPSR vs DSR

TCP over GPSR, AODV, DSR and DSDV

Speed(m/s)

Th

rou

gh

pu

t (K

bp

s)

Congestion Aware GPSR

Hot spot

Problem:Congestion area will cause long packet delay and

high loss probability Our approach:1. Go around the congestion area will decrease the delay,

but detour path is usually longer than the shortest path. Going through the long path will cause throughput loss.

2. Study packet delay, the tradeoff between congestion detour and throughput gain.

GPSR commentary• Very scalable:

– small per-node routing state – small routing protocol message complexity– robust packet delivery on densely deployed,

mobile wireless networks• TCP is extremely sensitive to path breakage

(timeout) -- It does very well with georouting• Outperforms DSR and AODV• Drawback: it requires knowledge of dest geo

coordinates (explicit forwarding node address)– Beaconing overhead– nodes may go to sleep (on and off) in

sensor networks

Geographic Random Forwarding (GeRaF)

- Forwarding in a Large Sensor Net• Nodes in turns go to sleep and wake

up, source does not know which nodes are on/off

• Source cannot explicitly address the next hop, must randomly select

• ideally, the best available node to act as a relay is chosen

• this selection is done a posteriori, i.e., after the transmission has taken place

• it is a receiver contention scheme

GeRaF: Key Idea

Goal: pick the relay closest to the destination broadcast message is sent, all active nodes within range receive it contention phase takes place: nodes closer to the destination are likely to win the winner becomes itself the source

Practical Implementation• major problem: how to pick the best relay?• solution: partition the area and pick relays

from slice closest to the destination• nodes can determine in which region they are• nodes in highest priority region contend first

Contention Resolution

• Assume 802.11 RTS/CTS

• Source transmits RTS with source and destination coordinates

• Stations in priority region #1 are solicited

• If none responds, stations in region #2 are solicited

Conclusions

• nodes who receive a message volunteer and contend to act as relays

• advantages: good for sensor net

– no need for complicated routing tables or routing-related signaling

– near-optimal multihop behavior, much better than alternative solutions (eg GAF, SPAN)

– significant energy/latency gains if nodes are densely deployed

Mobility assisted routing

• Mobility (of groups) was helpful to scale the routing protocol – see LANMAR

• Can mobility help in other cases?– Destination discovery (if coordinates not

know)– Mobility induced distributed

route/directory tree• Ref: H. Dubois Ferriere et al. ”Age Matters:

Efficient Route discovery in Mobile Ad Hoc Networks Using Encounter ages, Mobihoc 2003

Mobility Diffusion and “last encounter” routing

• Imagine a roaming node “sniffs” the neighborhood and learns/stores neighbors’ IDs

• Roaming node carries around the info about nodes it saw before

• Instead of searching for the destination, the source node searches for any intermediate node that encountered the destination more recently than did the source node itself.

• The intermediate node then searches for a node that encountered the destination yet more recently, and the procedure iterates until the destination is reached.

Mobility Diffusion and “last encounter” routing (cont.)

• If nodes move randomly and uniformly in the field (and the network is dense), there is a trail of nodes – like pointers – tracing back to each ID

• The superposition of these trails is a tree – it is a routing tree (to send messages back to source); or a distributed directory system (to map node ID to geo-coordinates, for example)

• “Last encounter” routing: next hop is the node that last saw the destination

Fresh algorithm – H. Dubois Ferriere, Mobihoc 2003

Mobility induced, distributed embedded route/directory tree

Benefits:

• (a) avoid overhead of periodic advertising of node location (eg, Landmark routing)

• (b) reduce flood search O/H (to find ID)

• (c ) avoid registration to location server (to DNS, say)

Issue:

• Motion pattern impact (localized vs random roaming)

“Direction” forwarding for mobile, large scale ad hoc networks

• In Distance Vector Routing (e.g., Bellman Ford, AODV etc.) node keeps pointer to “predecessor”

• When the predecessor moves, the path is broken • Alternate paths, even when available, are not used

Sink

Source

DV updatePredecessorData flow

Proposed solution: direction forwarding

Distance Vector not robust to mobility

Dest

Source

Primary PredecessorPrimary Path

Direction to Dest

Alternate Data Path

DV Update

Direction Forwarding

• Distance Vector update creates not only “predecessor”, but also “direction” entry

• Select “most productive” neighbor in forward direction

• If the network is reasonably dense, the path is salvaged

How to compute the “direction” Need “stable” local orientation system (say,

virtual compass) to determine direction of update Local (rather than global) reference is

required; Local reference system must be refreshed

fast enough to track avg local motion GPS will do (e.g., neighbors exchange (X, Y)

coordinates) If GPS not available, several non-GPS

coordinate systems have been recently published Sextant [Mobihoc ’05]; beacon DV; RFID’s

etc

Computing the “direction”(cont)

Compute “direction” to a destination when DV updates are received: If a DV update packet with a more recent

Seq # or smaller hop distance is received: New “direction” replaces the old one

The “direction” to the predecessor is used as the “direction” to the destination

If multiple DV updates received from different “predecessors” with same hop distance and seq # for the destination Take vector sum of directions

Computation of the “direction”

)(tan

)()(

12

121

212

212

XX

YY

YYXXr

Computation of the “direction”

Where the polar angle is the radian from the x-axis that is used as the direction of the predecessor node.

Suppose node A receives DV update packets from B & C

Compute the “directions” to predecessors node B & C, respectively,

A

C

B),( bb r

),( cc r

)1,( c

)1,( b

d

“Direction” to a destination

Unit vectors are used to combine the two “directions”

Directions to predecessors

Direction Forwarding vs Geo routing

• Geo-routing:– Direction points to destination– This direction may be unfeasible (holes, etc)– Global geo-coordinates (eg, GPS)– Geo Location Server– Robust to mobility

• Direction Forwarding– Direction of updates (always feasible)– Local (not global) position reference system– Advertisements from destination– Robust to mobility

Case study: apply Direct Forwarding (DFR) to LANMAR Routing

• LANMAR builds upon existing routing protocols– (1) “local ” routing algorithm that keeps

accurate routes within local scope < k hops (e.g., OLSR, FSR)

– (2) Landmark routes advertised to all mobiles using a Distance Vector approach

Logical GroupLogical Group

LandmarkLandmark

LANMAR +DFR

• LANMAR has proved to be very scalable to size

• However, as speed increases, performance degrades, even with group mobility!

• Problem was traced to failure of DV route advertising in high mobility

• We first tried to refresh more frequently: it did not work!

• Next step: try DFR

Simulation Experiments• Simulator: QualNet 3.8

– Standard IEEE 802.11 radio with a channel rate of 2Mbps and transmission range of 367 meters.

– Network field size: 2250m by 2250m• LANMAR is the protocol “hosting” DFR

– 225 nodes (or 360 nodes) equally distributed in 9 groups

– Mobility model: Group Mobility model• Traffic: CBR, 1 packets/sec, 512 bytes/packet

– The # of source-destination pairs is varied in the simulations to vary the offered traffic load

Performance as a function of speed

Delivery ratio vs. speed (Including packet loss due to disconnected

destination)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 2 4 6 8 10 12 14 16Mobility (m/sec)

Del

iver

y F

ract

ion

LANMAR(225 nodes)

DFR(225 nodes)

LANMAR(360 nodes)

DFR(360 nodes)

DFR

LANMAR

Conclusions and Future Work

• DFR: new forwarding strategy for table driven routing

• Direction Forwarding can improve LANMAR performance dramatically at high speeds

• Future Work:– Test DFR under local reference system– Apply DFR concept to AODV - Hybrid– TCP over {LANMAR, AODV} + DFR – Compare DFR with other backup route

schemes– Test DFR under more general mobility

models

Robust Ad Hoc Routing for Lossy Wireless Environment

• Challenges for routing in mobile ad hoc network

– Route breakage

– High BER

– Scalability

• The shortcomings of on-demand routing

• Not scalable for mobility

• The shortcomings of proactive routing

• Constant and high routing overhead

• The shortcomings of current Geo-routing

• Need Geo-Location Service, GLS

• “Face routing” is inefficient

Hybrid Routing: AODV-DFR

(AODV with Directional Forwarding Routing) • Combines on-demand and proactive routing

– When a source starts comm, it first finds the destination as in an on-demand fashion

– Once the destination is notified, it initiates periodic routing updates in a proactive fashion

• Utilizing an alternate path instantly based on “direction” to the destination if a path fails– resemblance with Georouting in the update

message– No location server system is required (not

like GPSR)

AODV-DFR

• Source initiates route discovery a la AODV – Destination, or any node that has a route,

replies– The path is set up

• Destination begins proactive advertisements (a la DV) after receiving data pkts from source – Intermediate nodes rebroadcast ads– Only for active connections– Period increases with distance from

destination (Fisheye concept)• Packet routing assisted by Direction Forward• The destination stops advertisement if it does

not receive pkts for some time

Performance Evaluation

• Compare AODV, AODV-DFR, GPSR and ADV (proactive and on-demand Hybrid Routing)– Performance: Delivery ratio, Packet delay,

Routing Overhead– Mobile & lossy network: UDP and TCP traffic

• Mobility Speed• Packet loss: uniformly distributed on a link

• Simulation– 100 nodes randomly moving in 1000x1000m– The traffic pairs are randomly distributed

over the network– UDP flows: pkt size 512 bytes, rate 1pkt/sec– TCP flows: NewReno, pkt size 1460 bytes

Mobile Network: Delivery Ratio

80 UDP flows

Mobile Network: Packet delay

80 UDP flows

Mobile Network: Routing Overhead

80 UDP flows

Mobile & Lossy Network: Delivery Ratio

UDP Flow number: 80 Mobility Speed: 10 m/s

Mobile & Lossy Network: Routing Overhead

UDP Flow number: 80 Mobility Speed: 10 m/s

TCP in Mobile Network

40 TCP flows

TCP in Mobile & Lossy Network

TCP flow number: 40 Mobility: 10 m/s

AODV-DFR Contributions

• A hybrid routing: proactive + on-demand• Robust to mobility and packet loss• Utilize location information for directional

forwarding with only local updates.• Low overhead• Provide better performance than AODV and

GPSR• Enhances AODV• Competitive with GPSR (does not require

“global” positioning such as GPS)• Ongoing work: local coordinate system;

integration of local and global coordinates (indoor+outdoor)