Outline
MANET overview Applications of MANET Issues in MANET MAC Protocols for MANET Routing Protocols for MANET Open issues and future directions
Mobile Ad Hoc Networks Formed by wireless
autonomous hosts Without (necessarily)
using a pre-existing infrastructure
Routes between hosts may potentially contain multiple hops
Host mobility cause route changes
Shared wireless channel
Why Ad Hoc Networks ? Ease of deployment
Speed of deployment
Decreased dependence on infrastructure
User flexibility
Application areas Military environments
Battle field: sensors, soldiers, vehicles Emergency operations
search-and-rescue policing and fire fighting
Civilian environments conference halls sports stadiums, Library, etc.
Personal area networking laptop, PDA, cell phone, ear phone, wrist watch
Challenges & Issues Medium Access Control
Distributed Operation Synchronization Hidden & Exposed terminal problem Access delay and Fairness Real Time Traffic support Low bandwidth Ease of snooping on wireless transmissions
Routing Mobility-induced route changes/packet losses High BER Location-dependent contention Looping Distributed Routing
Challenges & Issues Transport Layer Protocols
UDP – Highly Unreliable, may result into increasing congestion
TCP – Frequent path breaks, stale routing information, high error rate, frequent network partitions.
Energy Management Battery energy management Transmission Power management Processor and device power management
Security Denial of Service attack (DoS) Resource Consumption
Energy Depletion Buffer Overflow
Compromised Nodes Interference
Challenges & Issues Deployment Constraints
Environment Area of Coverage
Asymmetric Capabilities transmission ranges battery life processing capacity Speed/pattern of movement
MAC Protocols Goals
Operation should be distributed Support QoS for Real Time data Minimize Access Delay Fairness Scalable Power Control Mechnism Adaptive Rate Control Synchronization
MAC ProtocolsClassification of MAC Protocols Contention-Based Contention-Based with Reservation Contention-Based with Scheduling
Why is Routing in MANET Different? Host mobility
link failure/repair due to mobility different characteristics than those due to other
causes Rate of link failure/repair may be high when
nodes move fast Distributed Environment New performance criteria may be used
Route stability despite mobility Packet delivery ratio Routing Overhead
Routing Protocols in MANET Goals
Fully Distributed Adaptive to frequent topology changes Route computation and maintenance must
involve minimum number of nodes Routing state must be localized Must be loop free and free from stale route Converge must be quick Efficient resource utilization like BW, comp
power, battery, memory Provide QoS
Ad hoc Routing Protocols Proactive protocols (Table Driven)
(Eg.DSDV)
Reactive protocols (On-Demand)(Eg. AODV, DSR)
Hybrid protocols (ZRP, CEDAR)
Which approach achieves a better trade-off depends on the traffic and mobility patterns
Proactive Protocols Each node maintains a table having consistent,
up-to-date routing information from each node to every other node.
Respond to changes by propagating updates to throughout the network
Variations are on basis of number of tables required and the way updates are propagated.
Features: Traditional distributed shortest path routing protocols link-state or distance-vector protocol
Examples Destination-Sequenced Distance-Vector (DSDV) Clusterhead Gateway Switch Routing (CGSR) The Wireless Routing Protocol (WRP)
Reactive Protocols Creates route only when desired by the source node Source initiates a route discovery process within the
network Once a route has been established, it is maintained by
some form of route maintenance procedure Features:
Maintain routes only if needed Flooding of control message higher latency and lower overhead Source routing/hop-by-hop routing
Examples Ad hoc On Demand Distance Vector Protocol (AODV) Dynamic Source Routing Protocol (DSR) Temporally-Ordered Routing Algorithm (TORA) Associativity-Based Routing (ABR) Signal Stability Routing (SSR)
Hybrid Protocols
Hybrid routing protocols are proposed to combine the merits of both proactive and reactive routing protocols and overcome their shortcomings.
Features: Constrained link state maintenance Route established on-demand
Examples Zone Routing Protocol (ZRP) Core-Extraction Distributed Adhoc Routing (CEDAR)
DSDV Destination Sequenced Distance Vector
routing protocol Proactive Each node maintains its own sequences
number Updates (increments) at each change in
neighborhood information Used for loop freedom
Each node maintains routing table with entry for each node in the network
DSDV --- Routing Table at MN4
Dest Nexthop Metric DestSequenceMN1 MN2 2 406MN2 MN2 1 128MN3 MN2 2 564MN4 MN4 0 710MN5 MN6 2 392MN6 MN6 1 076MN7 MN6 2 128MN8 MN6 3 050
DSDV routing updates Each node periodically transmits updates
Includes its own sequences number, routing table updates
Nodes also send routing table updates for important link changes
When two routes to a destination received from two different neighbors Choose the one with greatest destination
sequence number If equal, choose the smaller metric (hop
count)
DSDV --- full dump
Full Dumps Carry all routing table information Transmitted relatively infrequently
Incremental updates Carry only information changed since
last full dump Fits within one network protocol data
unit If can’t, send full dump
DSDV --- link additions
When A joins network Node A transmits routing table: <A, 101, 0> Node B receives transmission, inserts <A, 101, A, 1> Node B propagates new route to neighbors <A, 101,
1> Neighbors update their routing tables: <A, 101, B, 2>
and continue propagation of information
DSDV --- link breaks
Link between B and D breaks Node B notices break
Update hop count for D and E to be infinity Increments sequence number for D and E
Node B sends updates with new route information <D, 203, infinite> <E, 156, infinite>
DSDV --- Summary Routes maintained through periodic and event
triggered routing table exchanges Incremental dumps and settling time used to
reduce control overhead Lower route request latency, but higher
overhead Perform best in network with low to moderate
mobility, few nodes and many data sessions Problems:
Not efficient for large ad-hoc networks Nodes need to maintain a complete list of routes.
Clusterhead Gateway Switch Routing (CGSR) Similar to DSDV except addressing being
employed and network organization Node are grouped into clusters. Clusterhead
selection algo (Least Cluster Change) is employed to elect a clusterhead
Gateways are used to relay packets between clusterheads.
Each node maintains cluster member table which stores destination cluster head for each node.
Also they maintain routing table, like DSDV to find next hop
CGSR--- Summary Easy to implement scheduling Better utilization of resources
Problems: Increase in path length Instability at high mobility
The Wireless Routing Protocol (WRP) Each node maintains 4 tables:
Distance table Routing table Link-cost table Message retransmission list (MRL)
DT contains matrix where each element contains distance and penultimate node reported by neighbor of a particular destination
MRL contains sequence no of update message, retransmission counter, ack required flag for each neighbor, list of update sent in update message
AODV The Ad-hoc On-Demand Distance Vector
Algorithm Reactive Pure on-demand route acquisition system Route discovery cycle used for route finding Maintenance of active routing Sequence number used for loop prevention and
route freshness criteria Descendant of DSDV Provides unicast and multicast communication
AODV --- Goal Quick adaptation under dynamic
link conditions Lower transmission latency Consume less network bandwidth
(less broadcast) Loop-free property Scalable to large network
AODV --- unicast route discovery
RREQ (route request) is broadcast Sequence Number:
Source SN: freshness on reverse route to source Destination SN: freshness on route to destination
RREQ message <bcast_id, dest_ip, dest_seqno, src_seqno,
hop_count> While forwarding, intermediate nodes record address of
neighbor from where first copy of RREQ is received, thus creating reverse path.
AODV --- unicast route discovery
RREP (route reply) is unicast back From destination if necessary From intermediate node if that node has a
recent route Intermediate node forwarding RREP stores this
info in their routing cache to set up a path to destination
Route timer is maintained with each entry. If idle for some time delete that route
As RREP is always forwarded over path of RREQ, it always expects symmetric links.
AODV --- route discovery (1)
1. Node S needs a route to D2. Create a route request (RREQ)
Enters D’s IP address, sequence number, S’s IP address, sequence number
Broadcasts RREQ to neighbors
AODV --- route discovery (2)
3. Node A receives RREQ Makes reverse route entry for S
Dest = S, nexthop = S, hopcount = 1 It has no route to D, so it broadcasts RREQ
4. Node C receives RREQ Makes reverse route entry for S
Dest = S, nexthop = A, hopcount = 2 It has route to D && seq# for route D > seq# in RREQ
Creates a route reply (RREP) Enters D’s IP address, sequence number, S’s IP address,
hopcount Unicasts RREP to A
AODV --- route discovery (3)
5. Node A receives RREP Unicasts RREP to S Makes forward route entry to D
Dest = D, nexthop = C hopcount = 2
6. Node S receives RREP Makes forward route entry to D
Dest = D, nexthop = A hopcount = 3 Sends data packets on route to D
AODV --- route maintenance (1)
Link between C and D breaks Node C invalidates route to D in routing table Node C creates route error (RERR) message
Lists all destinations which are now unreachable Sends to upstream neighbors
Node A receives RERR Checks whether C is its next hop on route to D Deletes route to D, and forwards RERR to S
AODV --- route maintenance (2)
Node S receives RERR Checks whether A is its next hop on route
to D Deletes route to D Rediscovers route if still needed
AODV --- Optimizations Expanding ring search
Prevents flooding of network during route discovery
Control Time to Live of RREQ Local repair
Repair breaks in active routes locally instead of notifying source
If first repair attempt is unsuccessful, send RERR to source
AODV --- Summary Reactive / On-demand Sequence numbers used for route
freshness and loop prevention Route discovery cycle Maintains only active routes Optimization can be used to reduce
overhead and increase scalability
Dynamic Source Routing (DSR) Two Phases
Route discovery Route maintenance
Before transmitting, a node consults its route cache. If unexpired route available, use that route else begin route discovery by broadcasting route request pkt.
RREQ contains add of Dest, Source Add, unique ID
Dynamic Source Routing (DSR) Node receiving the pkt checks, if it knows the
route to destination. If it does not then adds its own address to record route and forward the pkt to next neighbor.
RREP is generated by destination or any intermediate node knowing path to destination.
Responding node can use path to initiator for RREP if available else if symmetric links are supported, reverse path in record route.
If symmetric links are not supported, initiate RREQ with RREP being piggybacked
Dynamic Source Routing (DSR) Route Maintenance – Whenever a link
breakage is noticed, route error pkt is generated and broadcated and route containing that hop is deleted.
Apart from error pkt, ack is used to check correct operation of links (typically passive ack)
Associativity Based Routing (ABR) Route selection based stability of a route Stability is measured based count of beacons If a beacon is not received for some interval,
count is set to zero for a link A link is stable is it is leading to a stable
neighbor and same way unstable link
Associativity Based Routing (ABR) RREQ is flooded Intermediate nodes forwards RREQ by appending
its address and beacon count in it. When it reaches destination, dest waits for
TRouteSelectTime to receive more RREQ from different paths
Select route with maximum proportion of stable links
If two routes with same proportion of stable links, then choose shorter one.
But more priority is given to stability than length
Associativity Based Routing (ABR) Route Maintenance If a link is down, neighbor node detects the
breakage of link and initiates local recovery by broadcasting route repair pkt called Local Query(LQ) broadcast with limited TTL (Ex 2)
Advantages & Disadvantages Routes are stable so less chance of link failure Cons: Path may be longer Repetition of LQ
Hybrid Protocols Proactive protocol:
Pro-actively updates network state and maintains route regardless of whether any data traffic exists or not
Reactive protocol: Only determines route to a
destination if there is some data to be sent to the destination
Zone Routing Protocol(ZRP) Uses best features of reactive and proactive Uses proactive routing within a zone and reactive
outside the zone Intra Zone Routing Protocol (IARP) and Inter Zone
Routing Protocol (IERP) A routing zone of a given node is subset of nw
within which all nodes are reachable within less than or equal to zone radius hops
Interior nodes and peripheral nodes Each nodes maintains info abt all nodes in the
zone by periodic route update pkt (IARP)
Zone Routing Protocol(ZRP) When a pkt from s is to be transmitted to d S checks whether d is in its routing zone, if yes it
uses proactive routing table If no it broadcasts RREQ to all peripheral nodes If they know they respond with RREP else they also
forward to their peripherals until the destination is reached
All forwarding (RREQ) nodes appends their address to the RREQ to deliver RREP on that path
Broken link is handled by local repair and a path update message is sent to source
CEDAR Core-Extraction Distributed Ad Hoc Routing Dominator Set
Each node is in dominator sets or is the neighbor of one dominator node
Minimum Dominator Set and the links which length is no greater than 3 construct the core
Minimum Dominator Set and Core
Core Extraction Core extraction
Establishment & maintenance of a routing infrastructure called “core”
Finding core (Minimum Connected Dominating Sets) is NP-complete
Each node picks one core node as its dominator Dominator node is chosen based on degree of the
outgoing link Periodical Link state propagation
propagation of the link-state of stable high-bandwidth links in the core
Route Computation Route computation
route computation at the core nodes using all pair shortest path algorithm
S D
CEDAR --- Route Discovery
Node S informs its dominator core node A
Node A finds a route in the core network to the core node B which is the dominator for destination D
Core nodes on the above route between A and B then build a route from S to D using locally available link state information
CEDAR --- Summary Advantages
Route discovery/maintenance duties limited to a small number of core nodes
Link state propagation is a function of link stability/quality
Disadvantages Core nodes have to handle additional traffic,
associated with route discovery and maintenance
Hard to converge under high mobility
Special Constraints
Routing with special constrains Power Security QoS
Open issues and future directions
Power-Aware Routing: criteria
Define optimization criteria as a function of energy consumption. Examples:
Minimize energy consumed per packet
Minimize time to network partition due to energy depletion
Maximize duration before a node fails due to energy depletion
Power-Aware Routing: approach Assign a weight to each link
Weight of a link may be a function of energy consumed when transmitting a
packet on that link residual energy level
Prefer a route with the smallest aggregate weight
Security Issues in Mobile Ad Hoc Networks: What’s New ? Ad hoc network based on peer cooperation
Can you trust your peer? Wireless medium is easy to snoop on
Trace the path of active routes Easier for intruders to insert themselves into
the network Everybody is a “router” inject erroneous routing information divert network traffic, or make routing inefficient
Open Problems Address assignment problem
Stationary or auto-configuration? Improving interaction between protocol
layers Some routing protocol need feed back from MAC
to detect link status Position information from higher layer
Integration with Internet Existing ad hoc routing with infrastructure nodes Different network perspectives
Open ProblemsScalability
Short-range Throughput per node decreases at a rate 1/ , where N is the number of nodes This cannot be fixed except by physical layer improvements, such as directional antennas
Quality of service Need to provide best-effort service only for Voice, live video and file transfer
Client server model shift There is no server, but demand for basic services still exists. Address allocation, name resolution, authentication and service location are just examples of very basic services which are needed
Security Lack of any centralized network management or certification authority Networks are particularly prone to malicious behavior
Interoperation with the Internet Networks require some Internet connection Interface between the two are very different
Energy conservation Lifetime of a single battery and the whole network.
Node cooperation Why anyone should relay other people’s data
Interoperation What happens when two autonomous ad hoc networks move into same area