Upload
others
View
12
Download
0
Embed Size (px)
Citation preview
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 1 –
© James P.G. SterbenzITTC
James P.G. Sterbenz
Department of Electrical Engineering & Computer ScienceInformation Technology & Telecommunications Research Center
The University of Kansas
http://www.ittc.ku.edu/~jpgs/courses/mwnets
Mobile Wireless NetworkingThe University of Kansas EECS 882
MANET Routing
© 2004–2009 James P.G. Sterbenz29 October 2009 rev. 1.1
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-2
© James P.G. SterbenzITTC
Mobile Wireless NetworkingMANET Routing Algorithms and Protocols
MR.1 Routing algorithm alternativesMR.2 Example protocols
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 2 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-3
© James P.G. SterbenzITTC
MANET Algorithms and ProtocolsIntroduction
• Mobile ad hoc network (MANET)– mobile: node and groups of nodes move
• subject to a mobility model Lecture LM– wireless: mobility implies mostly wireless links
• weak, asymmetric, episodic connectivity Lecture PL– ad hoc: little or no reliance on network infrastructure
• from Latin: for this (purpose)
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-4
© James P.G. SterbenzITTC
MANET RoutingMR.1 Routing Algorithm Alternatives
MR.1 Routing algorithm alternativesMR.2 Example protocols
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 3 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-5
© James P.G. SterbenzITTC
MANET RoutingChallenges
• Routing algorithmpurpose and distinction from forwarding?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-6
© James P.G. SterbenzITTC
MANET RoutingChallenges
• Routing algorithm discovers path EECS780 lecture NR– between source(s) and destination(s)
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 4 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-7
© James P.G. SterbenzITTC
MANET RoutingChallenges
• Routing algorithm discovers path– between source(s) and destination(s)
• Challenges in MANETs– episodic connectivity and mobility
implications?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-8
© James P.G. SterbenzITTC
MANET RoutingChallenges
• Routing algorithm discovers path– between source(s) and destination(s)
• Challenges in MANETs– episodic connectivity and mobility– topology and link state keeps changing– difficult or impossible to maintain consistent information
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 5 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-9
© James P.G. SterbenzITTC
MANET RoutingChallenges
• Routing algorithm discovers path– between source(s) and destination(s)
• Challenges in MANETs– episodic connectivity and mobility– topology and link state keeps changing– difficult or impossible to maintain consistent information
• Conventional routing algorithms do not work wellwhy?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-10
© James P.G. SterbenzITTC
MANET RoutingChallenges
• Routing algorithm discovers path– between source(s) and destination(s)
• Challenges in MANETs– episodic connectivity and mobility– topology and link state keeps changing– difficult or impossible to maintain consistent information
• Conventional routing algorithms do not work well– assume relatively stable topologies and link state– convergence difficult or impossible
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 6 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-11
© James P.G. SterbenzITTC
MANET RoutingAlgorithm vs. Protocol
• Routing algorithm– algorithmic formalism to describe path discovery
example: link state / Dijkstra shortest path algorithm
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-12
© James P.G. SterbenzITTC
MANET RoutingAlgorithm vs. Protocol
• Routing algorithm– algorithmic formalism to describe path discovery
• Routing protocol– specification of algorithm as state machines and PDUs– permits interoperable operation among nodes
example: OSPF protocol using link state & Dijkstra– common to be sloppy about algorithm/protocol difference
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 7 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-13
© James P.G. SterbenzITTC
MANET RoutingAlgorithm vs. Protocol
• Routing algorithm– algorithmic formalism to describe path discovery
• Routing protocol– specification of algorithm as state machines and PDUs– permits interoperable operation among nodes
• Routing Implementation– software or hardware implementation of protocol– code and data structures compliant to protocol specification
example: OSPF protocol implementation in Cisco IOS
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-14
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives
• Routing algorithm type• Topological structure• Routing update mechanism• Temporal information• Resource and context awareness
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 8 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-15
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Algorithm Type
• Routing algorithm type– distance vector– link state– source route
• Topological structure• Routing update mechanism• Resource and context awareness
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-16
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Distance Vector
• Distance vectoroperation and relative advantages?
• Link state• Source route
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 9 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-17
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Distance Vector
• Distance vector– nodes maintain vectors of shortest next hops per destination– Bellman-Ford path computation– minimises number of hops– slow to converge for large networks– subject to routing loops and count-to-infinity– only appropriate for small networks
(wired ) examples?
• Link state• Source route
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-18
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Distance Vector
• Distance vector– nodes maintain vectors of shortest next hops per destination– Bellman-Ford path computation– minimises number of hops– slow to converge for large networks– subject to routing loops and count-to-infinity– only appropriate for small networks– (wired) examples: RIP, IGRP, EIGRP
• Link state• Source route
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 10 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-19
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Link State
• Distance vector• Link state
operation and relative advantages?
• Source route
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-20
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Link State
• Distance vector• Link state
– every node maintains complete topology database– Dijkstra shortest path algorithm– flooding of link state advertisements (LSAs)– rapid convergence– better for large networks than distance vector
(wired ) examples?
• Source route
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 11 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-21
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Link State
• Distance vector• Link state
– every node maintains complete topology database– Dijkstra shortest path algorithm– flooding of link state advertisements (LSAs)– rapid convergence– better for large networks than distance vector– (wired) examples: OSPF, ISIS
• Source route
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-22
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Source Routed
• Distance vector• Link state• Source route
operation and relative advantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 12 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-23
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Source Routed
• Distance vector• Link state• Source route
– source constructs the sequence of nodes to destination
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-24
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Source Routed
• Distance vector• Link state• Source route
– source constructs the sequence of nodes to destination– route discovery algorithm needed– state carried in packet headers
– rather than maintained in forwarding tables
(wired ) examples?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 13 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-25
© James P.G. SterbenzITTC
MANET Routing AlgorithmsAlgorithm Type: Source Routed
• Distance vector• Link state• Source route
– source constructs the sequence of nodes to destination– route discovery algorithm needed– state carried in packet headers
– rather than maintained in forwarding tables
– (wired) examples: IP source route option (rarely used)future Internet proposals: NewArch, PoMo
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-26
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Topological Structure
• Routing algorithm type• Topological structure
– flat– hierarchical
• Routing update mechanism• Resource and context awareness
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 14 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-27
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Flat
• Flat– all nodes have a uniform view of identifier space
limitations?
• Hierarchical
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-28
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Flat
• Flat– all nodes have a uniform view of identifier space– appropriate for small networks
(wired ) examples?
• Hierarchical
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 15 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-29
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Flat
• Flat– all nodes have a uniform view of identifier space– appropriate for small networks– (wired) examples: RIP, IGRP
• Hierarchical
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-30
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Hierarchical
• Flat• Hierarchical
motivation?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 16 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-31
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Hierarchical
• Flat• Hierarchical
– allows scalable construction of large networks• hierarchy is almost always the first answer to network scaling
implementation?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-32
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Hierarchical
• Flat• Hierarchical
– allows scalable construction of large networks• hierarchy is almost always the first answer to network scaling
– nodes divided into clusters– each cluster has uniform access to other members
• e.g. LSA flooding and topology databases
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 17 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-33
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Hierarchical
• Flat• Hierarchical
– allows scalable construction of large networks• hierarchy is almost always the first answer to network scaling
– nodes divided into clusters– each cluster has uniform access to other members
• e.g. for link state advertisement flooding and topo databases
– clusters abstracted into virtual nodes at next level
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-34
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Hierarchical
• Flat• Hierarchical
– allows scalable construction of large networks• hierarchy is almost always the first answer to network scaling
– nodes divided into clusters– each cluster has uniform access to other members
• e.g. for link state advertisement flooding and topo databases
– clusters abstracted into virtual nodes at next level(wired ) examples?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 18 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-35
© James P.G. SterbenzITTC
MANET Routing AlgorithmsTopological Structure: Hierarchical
• Flat• Hierarchical
– allows scalable construction of large networks• hierarchy is almost always the first answer to network scaling
– nodes divided into clusters– each cluster has uniform access to other members
• e.g. for link state advertisement flooding and topo databases
– clusters abstracted into virtual nodes at next level– (wired) examples: OSPF (2-level), P-NNI (n -level)
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-36
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Update Mechanism
• Routing algorithm type• Topological structure• Routing update mechanism
– proactive or table-driven– reactive or on-demand– predictive
• Resource and context awareness
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 19 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-37
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Proactive / Table-Driven
• Proactive or table-driven– compute paths that may be needed (proactive )– insert into forwarding tables (table-driven )
sound familiar?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-38
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Proactive / Table-Driven
• Proactive or table-driven– compute paths that may be needed (proactive )– insert into forwarding tables (table-driven )– conventional routing technique used in Internet
advantages and disadvantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 20 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-39
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Proactive / Table-Driven
• Proactive or table-driven– compute paths that may be needed (proactive )– insert into forwarding tables (table-driven ) – conventional routing technique used in Internet+ low communication startup latency– overhead of continuous route maintenance
• routing must reconverge in response to topology changes
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-40
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
difference from proactive?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 21 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-41
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-42
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing– requires route discovery signalling
advantages and disadvantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 22 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-43
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing– requires route discovery signalling+ lower overall overhead
assumption?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-44
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing– requires route discovery signalling+ lower overall overhead
• assuming new route request rate not very high
– higher startup delaywhy?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 23 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-45
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing– requires route discovery signalling+ lower overall overhead
• assuming new route request rate not very high
– higher startup delay• each communication flow must wait for route discovery
mitigation?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-46
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing– requires route discovery signalling+ lower overall overhead
• assuming new route request rate not very high
– higher startup delay• each communication flow must wait for route discovery• unless route cached from recent flow between same node pair
– overhead of route maintenancewhy?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 24 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-47
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Reactive / On-Demand
• Proactive or table-driven• Reactive or on-demand
– compute paths only when needed (reactive on-demand )– may insert into forwarding tables or drive source routing– requires route discovery signalling+ lower overall overhead
• assuming new route request rate not very high
– higher startup delay• each communication flow must wait for route discovery• unless route cached from recent flow between same node pair
– overhead of route maintenance• signalling to alter routes with topology changes if rate high
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-48
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive
• Proactive or table-driven• Reactive or on-demand• Predictive
what and why?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 25 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-49
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive
• Proactive or table-driven• Reactive or on-demand• Predictive
– predict where packets go in advancewhy?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-50
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive
• Proactive or table-driven• Reactive or on-demand• Predictive
– predict where packets go in advance– necessary when mobility exceeds ability to converge– how?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 26 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-51
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive
• Proactive or table-driven• Reactive or on-demand• Predictive
– predict where packets go in advance– necessary when mobility exceeds ability to converge– location management with trajectory prediction
• exploit knowledge about mobility model
lecture LM
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-52
© James P.G. SterbenzITTC
• Routing and forwarding expect mobility• Use location/trajectory information where available
– unicast when predictable (e.g. planetary or racetrack UAV)
MANET Routing AlgorithmsUpdate Mechanism: Predictive
1
destination
source
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 27 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-53
© James P.G. SterbenzITTC
• Routing and forwarding expect mobility• Use location/trajectory information where available
– unicast when predictable (e.g. planetary or racetrack UAV)
MANET Routing AlgorithmsUpdate Mechanism: Predictive
2
destination
source
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-54
© James P.G. SterbenzITTC
• Routing and forwarding expect mobility• Use location/trajectory information where available
– unicast when predictable (e.g. planetary or racetrack UAV)– multicast to area of expected location (spray routing)
MANET Routing AlgorithmsUpdate Mechanism: Predictive
3
destination
source
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 28 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-55
© James P.G. SterbenzITTC
• Routing and forwarding expect mobility• Use location/trajectory information where available
– unicast when predictable (e.g. planetary or racetrack UAV)– multicast to area of expected location (spray routing)
MANET Routing AlgorithmsUpdate Mechanism: Predictive
4
destination
source
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-56
© James P.G. SterbenzITTC
• Routing and forwarding expect mobility• Use location/trajectory information where available
– unicast when predictable (e.g. planetary or racetrack UAV)– multicast to area of expected location (spray routing)
• cluster may have inherent broadcast or epidemic routing
MANET Routing AlgorithmsUpdate Mechanism: Predictive
5
destination
source
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 29 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-57
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive Scenario
• Very high relative velocity– Mach 7 ≈ 10 s contact– dynamic topology
• Communication channel– limited spectrum– asymmetric links
• data down omni• C&C up directional
• Multihop– among TAs– through relay nodes
GSGS
RN
TATA
TAs
Internet
GWGW
AN – airborne nodeRN – relay node
GS – ground stationGW – gateway
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-58
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive Scenario
15Mach 7.010TA – TA
300Mach 3.5100GS – TA
Single-Hop Worst Case
135800 knots15TA – TA
2520400 knots140GS – TA
Single-Hop Best Case
Contact Duration [sec]
Relative VelocityTransmit Range [nmi]Scenario
• Multihop case significantly harder– probability of stable end-to-end path very low
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 30 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-59
© James P.G. SterbenzITTC
MANET Routing AlgorithmsUpdate Mechanism: Predictive Example
• AeroRP airborne routing protocol– [Jabbar-Sterbenz-2009]
• Proactive with limited updates– eliminates delay of reactive – while limiting overhead of proactive
• Exploiting cross-layer information– explicit cross-layering provided by AeroNP– predictive using geolocation and trajectory information
• Snooping to assist neighbor discovery– overheard transmissions indicate neighbor presence
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-60
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Resource / Context Aware
• Routing algorithm type• Topological structure• Routing update mechanism• Resource and context awareness
– power- and energy-aware– geographical– communication environment
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 31 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-61
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Resource Aware
• Resource and context awareness– consider resources in routing algorithm– generally to conserve battery usage lecture EM
• processing, memory, bandwidth, transmission power• remaining battery life
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-62
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Context Aware
• Resource and context awareness– consider context of node in routing algorithm– geographical location-based routing
• geolocation as addressing (e.g. sensor networks) lecture SN• motion trajectories for high-mobility scenarios
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 32 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-63
© James P.G. SterbenzITTC
MANET Routing AlgorithmsDesign Alternatives: Context Aware
• Resource and context awareness– consider context of node in routing algorithm– communications environment
• disruption-tolerant routing lecture RS
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-64
© James P.G. SterbenzITTC
MANET RoutingMR.2 Example Protocols
MR.1 Routing algorithm alternativesMR.2 Example protocols
MR.2.1 DSDVMR.2.2 AODVMR.2.3 DSRMR.2.4 ZRPMR.2.5 OLSR
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 33 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-65
© James P.G. SterbenzITTC
MANET Routing AlgorithmsSimple Algorithm
What is the simplest possible routing algorithm?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-66
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
15
21
3
6 54
7
108
91112
1314
0
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 34 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-67
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
15
21
3
6 54
7
108
91112
1314
1
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-68
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
15
21
3
6 54
7
108
91112
1314
2
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 35 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-69
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
15
21
3
6 54
7
108
91112
1314
3
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-70
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
15
21
3
6 54
7
108
91112
1314
4
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 36 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-71
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
15
21
3
6 54
7
108
91112
1314
5
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-72
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding:send all packets to all nodes– simplest possible
routing algorithm
Advantages?
15
21
3
6 54
7
108
91112
1314
6
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 37 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-73
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding: send all packets to all nodes• Advantages:
+ simplest possible routing algorithm+ use sequence numbers to avoid forwarding duplicates
• each node discards an incoming packet already forwarded
+ may increase probability of deliverywhy?
Disadvantages?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-74
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding: send all packets to all nodes• Advantages:
+ simplest possible routing algorithm+ use sequence numbers to avoid forwarding duplicates
• each node discards an incoming packet already forwarded
+ may increase probability of delivery• every possible path explored
Disadvantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 38 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-75
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding: send all packets to all nodes• Advantages:
+ simplest possible routing algorithm+ use sequence numbers to avoid forwarding duplicates+ may increase probability of delivery
• Disadvantages– many packets sent that are not needed– bandwidth intensive in a bandwidth constrained environment– may reduce probability of delivery
why?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-76
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding
• Flooding: send all packets to all nodes• Advantages:
+ simplest possible routing algorithm+ use sequence numbers to avoid forwarding duplicates+ may increase probability of delivery
• Disadvantages– many packets sent that are not needed– bandwidth intensive in a bandwidth constrained environment– may reduce probability of delivery
• congestion and collisions
Alternative?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 39 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-77
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding Alternatives
• Alternative to flooding– use routing algorithm to find possible path– forward data packets directly along that path
How to find possible paths?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-78
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding Alternatives
• Alternative to flooding– use routing algorithm to find possible path– forward data packets directly along that path
• To find possible paths– flood control messages: route discovery
how is this different?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 40 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-79
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding Alternatives
• Alternative to flooding– use routing algorithm to find possible path– forward data packets directly along that path
• To find possible paths– flood control messages
• still flooding, but much lower overhead than every data packet• reuse path for packets in flow• cache route for subsequent flows
– exploit other knowledge• location management so sender can efficiently track receiver
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-80
© James P.G. SterbenzITTC
MANET Routing AlgorithmsFlooding Alternatives
• Alternative to flooding– use routing algorithm to find possible path– send data packets directly along that path
• To find possible paths– flood control messages
• still flooding, but much lower overhead than every data packet• reuse path for packets in flow• cache route for subsequent flows
– exploit other knowledge• location management so sender can efficiently track receiver
• There are many proposed MANET routing protocols
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 41 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-81
© James P.G. SterbenzITTC
MANET Routing ProtocolsExamples: IETF Experimental
• There are many proposed MANET routing protocols• A few are part of IETF MANET working group
http://www.ietf.org/html.charters/manet‐charter.html
– currently all RFCs are “experimental” status• increases chance of implementation in products• RFC 3561: AODV (ad hoc on-demand distance vector protocol)• RFC 3626: OLSR (optimized link state routing protocol)• RFC 3684: TBRPF (topology broadcast based on RPF)
– RPF (reverse path forwarding)
• RFC 4728: DSR (dynamic source routing protocol)
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-82
© James P.G. SterbenzITTC
MANET Routing ProtocolsExamples: Current IETF Internet Drafts
• There are many proposed MANET routing protocols• A few are part of IETF MANET working group
– currently all RFCs are “experimental” status– some in proposals are currently in Internet Draft form
• DYMO (dynamic MANET on-demand routing) – AODV successor• OLSR version 2• SMF (simplified multicast forwarding)
– IDs (Internet drafts)• can be part of IETF working group agenda• can be sponsored by individuals
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 42 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-83
© James P.G. SterbenzITTC
MANET Routing ProtocolsExamples: Dead IETF Internet Drafts
• There are many proposed MANET routing protocols• A few are part of IETF MANET working group
– currently all RFCs are “experimental” status– some in proposals are currently in Internet Draft form– some proposals never became RFCs
• ABR (associativity based routing)• MM (mobile mesh)• TORA (temporally ordered routing algorithm)• ZRP (zone routing protocol)
etc.
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-84
© James P.G. SterbenzITTC
MANET Routing ProtocolsExamples: Other Proposals
• There are many proposed MANET routing protocols• A few are part of IETF MANET working group• Many more research proposals
– DSDV (destination sequenced distance vector)– WRP (wireless routing protocol)
etc.– some designed for very specialised domains
• tactical military networks• vehicles and aircraft (e.g. AeroRP)• disruption-tolerant networks (e.g. P-WARP)
– IETF RFC status ≠ probability of deployment
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 43 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-85
© James P.G. SterbenzITTC
MANET Routing ProtocolsExamples
• Each proposed protocol could be an entire lecture– we will only provide an overview of a selection– somewhat arbitrary choice
• based on historical significance• based on how well known
– remember: none of these have seen significant deployment
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-86
© James P.G. SterbenzITTC
MANET Routing ProtocolsExamples
noneproactiveflatlink stateOLSR
nonehybridflatvariesZRP
nonereactiveflatsource routing
DSR
nonereactiveflatdistance vector
AODV
noneproactiveflatdistance vector
DSDV
Resource Context
Update Mechanism
Topological Structure
Routing AlgorithmProtocol
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 44 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-87
© James P.G. SterbenzITTC
MANET Routing ProtocolsNetwork Layer Functionality
• Addressing: node identifiers• Routing: path discovery• Forwarding: next-hop decision and transfer• Signalling: MANET control messages• Traffic management
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-88
© James P.G. SterbenzITTC
MANET Routing ProtocolsAddressing Alternatives
• Addressing: node identifiersalternative possibilities?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 45 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-89
© James P.G. SterbenzITTC
MANET Routing ProtocolsAddressing Alternatives
• Addressing: node identifiers– alternative possibilities
• IEEE 48b MAC addresses– de facto standard for unique network interface addressing– much larger addresses than needed for typical MANET
• IPv4 of IPv6 addresses– de facto standard for network addresses– useful when MANET connected to Internet
• MANET protocol-specific identifiers– flat: address space related to maximum size of MANET– hierarchical: size relate to cluster size and number of levels
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-90
© James P.G. SterbenzITTC
MANET Routing ProtocolsImplementation Alternatives
• Implementation alternatives – standalone– IP integration– IP encapsulation– UDP/IP encapsulation
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 46 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-91
© James P.G. SterbenzITTC
MANET Routing ProtocolsImplementation Options: Standalone
• Standalone– MANET packet (encapsulated in a link/MAC frame)– no dependencies on TCP/IP
• Node addresses in MANET header• Alternative address schemes:
– protocol-specific identifiers (flat or hierarchical)– IPv4, IPv6, or IEEE 48b MAC addresses can still be used
• Payload (determined by header bit) consists of either– MANET signalling message– data
link payloadMANET
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-92
© James P.G. SterbenzITTC
MANET Routing ProtocolsImplementation Options: IP Integration
• IP integration– MANET packet (encapsulated in a link/MAC frame)
• Node addresses in IP header (orginating, target)– other IP fields used such as TTL– extended by MANET subheader– additional addresses and control fields
• Payload consist of either– MANET signalling message
• Example: DSR (IP protocol ID=48)– MANET subheader contains protocol ID of payload
link payloadIP MANET
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 47 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-93
© James P.G. SterbenzITTC
MANET Routing ProtocolsImplementation Options: IP Encapsulation
• MANET signalling message encapsulated in IP– IP protocol number defines MANET protocol
• MANET data messages conventional UDP– using IP source = origin node, destinaton = target node
• IP protocol IDs– 138: generic MANET
IPlink payloadMANET138
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-94
© James P.G. SterbenzITTC
MANET Routing ProtocolsImplementation Options: UDP Encapsulation
• MANET signalling message encapsulated in UDP– UDP port number defines MANET protocol
• MANET data messages conventional UDP– using IP source = origin node, destinaton = target node
• Examples and UDP ports– 269: generic MANET– 654: AODV– 698: OLDR
IP UDPlink payloadMANET26917
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 48 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-95
© James P.G. SterbenzITTC
MANET Routing ProtocolsMR.2.2 DSDV
MR.1 Routing algorithm alternativesMR.2 Example protocols
MR.2.1 DSDVMR.2.2 AODVMR.2.3 DSRMR.2.4 ZRPMR.2.5 OLSR
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-96
© James P.G. SterbenzITTC
DSDVOverview
• DSDV: destination sequenced distance vector– one of the first MANET routing protocols [Perkins 1994]
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 49 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-97
© James P.G. SterbenzITTC
DSDVOverview
• DSDV: destination sequenced distance vector– one of the first MANET routing protocols [Perkins 1994]
• Table driven / proactive– each node maintains table– row for each possible destination
• next hop to reach destination• number of hops to destination
– sequence number to know which is most recent update
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-98
© James P.G. SterbenzITTC
DSDVOverview
• DSDV: destination sequenced distance vector– one of the first MANET routing protocols [Perkins 1994]
• Table driven / proactive• Updates exchanged between immediate neighbours
– single update when topology change at a given node– full dump when significant topology change at a given node– periodic synchronisation
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 50 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-99
© James P.G. SterbenzITTC
DSDVExample Network and Table
15
21
3
6 54
7
108
91112
1314
256451521436141984513190351217636111422610186429170358162327144166134155322542622322122
Seq#DistNextDest
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-100
© James P.G. SterbenzITTC
DSDVExample Network and Table
15
21
3
6 54
7
108
91112
1314
256451521436141984513190351217636111422610186429170358162327144166134155322542622322122
Seq#DistNextDest
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 51 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-101
© James P.G. SterbenzITTC
DSDVExample Network and Table
15
21
3
6 54
7
108
91112
1314
256451521436141984513190351217636111422610186429170358162327144166134155322542622322122
Seq#DistNextDest
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-102
© James P.G. SterbenzITTC
DSDVExample Network and Table
15
21
3
6 54
7
108
9
11
12
1314
256451521436141984513190351218045111422610186429170358162327144166134155322542622322122
Seq#DistNextDest
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 52 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-103
© James P.G. SterbenzITTC
DSDVAdvantages and Disadvantages
• Advantages– minor adaptation of wired distance vector protocol
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-104
© James P.G. SterbenzITTC
DSDVAdvantages and Disadvantages
• Advantages– minor adaptation of wired distance vector protocol
• Disadvantages– high overhead with non-trivial changes
• mobility• episodic link connectivity
– stale information before updates propagate• packets may be forwarded along wrong path
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 53 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-105
© James P.G. SterbenzITTC
WRPOverview
• WRP: wireless routing protocol– one of the first MANET routing protocols [Murthy, JJ 1996]– similar in concept to DSDV
• Table driven / proactive– each node maintains table– multiple tables for more accurate information
• distance• routing: distance, predecessor, successor, status• link cost• message retransmission list
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-106
© James P.G. SterbenzITTC
MANET Routing ProtocolsMR.2.2 AODV
MR.1 Routing algorithm alternativesMR.2 Example protocols
MR.2.1 DSDVMR.2.2 AODVMR.2.3 DSRMR.2.4 ZRPMR.2.5 OLSR
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 54 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-107
© James P.G. SterbenzITTC
AODVOverview
• AODV (ad hoc on-demand distance vector)– early MANET routing protocol [Perkins, Belding-Royer 1999]– adopted by IETF MANET working group [RFC 3561]
• On demand / reactive successor to DSDV– each node has forwarding table– paths only included when needed (reactive)
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-108
© James P.G. SterbenzITTC
AODVMessage Types
• AODV packets are for control only– encapsulate in UDP in IP in MAC frame– UDP port 654 for AODV
• Message types (AODV is only a routing protocol)1 = RREQ route request to discover path2 = RREP route reply to confirm path3 = RERR route error when link goes down4 = RREP‐ACK route reply ACK
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 55 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-109
© James P.G. SterbenzITTC
AODV OperationRoute Discovery
• Flood route request: RREQ message path not known– using broadcast IP address 255.255.255.255– IP TTL determines scope of flood
• Route discovery occurs when:– destination node previously unknown to node– previously valid route expires– previously valid route marked as invalid by RERR
• RREQ for destination node by originating node
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-110
© James P.G. SterbenzITTC
AODV Route DiscoveryRREQ Packet Format1
• RREQ: type = 01• Flags
– J: join (multicast)– R: repair (multicast)– G: gratuitous RREP
should be unicast to destination
– D: destination onlycan reply to this RREQ
– U: unknown destination seq #
– remaining bits reserved
type = 01
destination IP address
destination sequence number
originator IP address
originator sequence number
RREQ ID
hop countJRGDU
24B
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 56 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-111
© James P.G. SterbenzITTC
AODV Route DiscoveryRREQ Packet Format2
• RREQ: type = 01• Flags• Hop count
– number of hops from originator to this node
• RREQ ID– identifier of RREQ for
given originator– in combination with IP
address unique ID foreach RREQ
type = 01
destination IP address
destination sequence number
originator IP address
originator sequence number
RREQ ID
hop countJRGDU
24B
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-112
© James P.G. SterbenzITTC
AODV Route DiscoveryOperation: RREQ Packet Format3
• RREQ: type = 01• Flags• Hop count• RREQ ID• Destination IP address
of desired path• Destination sequence number• Originator IP address of RREQ• Originator sequence number
type = 01
destination IP address
destination sequence number
originator IP address
originator sequence number
24B
hop count
RREQ ID
JRGDU
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 57 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-113
© James P.G. SterbenzITTC
AODV OperationRoute Discovery: RREQ Flood
• Determine sequence numbers and ID – dest seq# is last used from routing table or set U flag if none– orig seq#– RREQ ID is incremented from last used by originator
• Rate– rate of RREQ messages limited to RREQ_RATELIMIT
• default = 10 msg/sec
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-114
© James P.G. SterbenzITTC
AODVOperation: Route Discovery
• Every intermediate node stores reverse path– bidirectional links required
• RREP returned to originator IP address when– fresh route located in intermediate node– destination receives the first RREQ from a given originator– this constructs distance vector (not Bellman-Ford)
• RREP uses reverse pointers in each node– establishes predecessor and successor nodes for a flow
• Forwarding table entries used by multiple flows– soft state: entries time out
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 58 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-115
© James P.G. SterbenzITTC
AODV Route DiscoveryRREP Packet Format1
• RREP: type = 02• Flags
– R: repair (multicast)– A: ACK required
• Prefix size• Hop count
• # hops orig→dest
• Lifetime• timer to prevent stale
RREPs from being used
type = 02
destination IP address
destination sequence number
originator IP address
lifetime
hop countRA
20B
prfix sz
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-116
© James P.G. SterbenzITTC
AODVOperation: Route Maintenance
• Adjacent nodes exchange periodic HELLO messages• Link failure detected
– forwarding failure– timeout of HELLO messages– failure to receive MAC-layer ACKs
• Route error (RERR) messages to neighbours– sequence numbers avoid broken paths and loops
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 59 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-117
© James P.G. SterbenzITTC
AODVOperation: Data Transfer
• Data packets forwarded hop-by-hop to destination– each hop uses forwarding table to lookup next hop
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-118
© James P.G. SterbenzITTC
AODVAdvantages and Disadvantages
Advantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 60 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-119
© James P.G. SterbenzITTC
AODVAdvantages
• Advantages+ relatively simple algorithm+ reduces flooding of control messages
• but flooded for every source–destination pair
+ paths maintained only for needed routes+ soft state allows amortisation over…
• time: multiple flows
+ good for long flows in a relatively stable network
Disadvantages?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-120
© James P.G. SterbenzITTC
AODVDisadvantages
• Disadvantages– mobility and episodic connectivity significantly impact– flooding of RREQ control messages
• may reach all nodes in a large network• required for every source–destination pair
– stale forwarding table entries lead to inconsistent routes– overhead of HELLO messages
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 61 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-121
© James P.G. SterbenzITTC
AODVDeployment
• 802.11s uses HWMP (hybrid wireless mesh protocol)– based in part on AODV
why is this a good choice for 802.11s mesh networking?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-122
© James P.G. SterbenzITTC
AODVDeployment
• 802.11s uses HWMP (hybrid wireless mesh protocol)– based in part on AODV– AODV: self-organising– AODV: occasional topology changes– AODV: small networks
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 62 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-123
© James P.G. SterbenzITTC
MANET Routing ProtocolsMR.2.3 DSR
MR.1 Routing algorithm alternativesMR.2 Example protocols
MR.2.1 DSDVMR.2.2 AODVMR.2.3 DSRMR.2.4 ZRPMR.2.5 OLSR
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-124
© James P.G. SterbenzITTC
DSROverview
• DSR (dynamic source routing)– early MANET routing protocol [Johnson, Maltz 1996]– adopted by IETF MANET working group [RFC 4728]
• On demand / reactive– source routing
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 63 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-125
© James P.G. SterbenzITTC
20B
DSRPacket Format: IP Header
• IP header– prot id = 48– IHL = 20– source addr [32b]
source node address– destination addr [32b]
controls dissemination– TTL [ 8b]
limits scope of flooding
• Fixed option hdr [32b]• Option
fixed option header4B
option
04 total length20 TOS
fragment id
TTL prot=48 header checksum
source address
destination address
frag offsetDF
MF0
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-126
© James P.G. SterbenzITTC
20B
DSRPacket Format: Fixed Option Header
• IP header• Fixed option header [32b]
– next hdr [ 8b]IP protocol idof encapsulated transport
– F [ 1b]flow state header flag
– payld len [16b]length DSR options hdr
• Option [length]
next hdr payload length4B resvF
option
04 total length20 TOS
fragment id
TTL prot=48 header checksum
source address
destination address
frag offsetDF
MF0
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 64 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-127
© James P.G. SterbenzITTC
DSROption Types
• DSR options are shim header following IP header– IP protocol ID = 48 reserved for DSR
• Option types (DSR is only a routing protocol)1 = RREQ route request path discovery2 = RREP route reply path establishment3 = RERR route error
160 = acknowledgement request32 = acknowledgement96 = DSR source route option
224 = Pad1 option0 = PadN option
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-128
© James P.G. SterbenzITTC
DSROperation: Route Discovery
• When source does not know path to destination:• Flood route request (RREQ)
– every node appends its 32-bit address to RREQ– constructs source route while packet is travelling to target
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 65 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-129
© James P.G. SterbenzITTC
24B
DSRPacket Format: RREQ Packet1
• IP header [20B]– source addr [32b]
originating address– destination addr [32b]
IP broadcast to flood 255.255.255.255
– TTL = [1…255] [ 8b]controls scope of flood
• Fixed option hdr [ 4B]• Option [8+4nB]
04 total length20 TOS
fragment id
TTL prot=48 header checksum
source address = orig. address
255.255.255.255
frag offsetDF
MF0
nxt=59 payload lengthresvF
identificationtype=1 opt len
target address
address[1]
address[1]
address[n]
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-130
© James P.G. SterbenzITTC
24B
DSRPacket Format: RREQ Packet2
• IP header [20B]• Fixed option hdr [ 4B]• Option [8+4nB]
– type = 1 RREQ [ 8b]– unique pkt id [16b]– target address [32b]– addresses [32b]
each node adds its address to this list to form source route
04 total length20 TOS
fragment id
TTL prot=48 header checksum
source address = orig. address
255.255.255.255
frag offsetDF
MF0
nxt=59 payload lengthresvF
identificationtype=1 opt len
target address
address[1]
address[1]
address[n]
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 66 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-131
© James P.G. SterbenzITTC
DSROperation: Route Discovery
• When source does not know path to destination:• Flood route request (RREQ)
– every node appends its ID to RREQ– constructs source route while packet is travelling to target
• Target sends route reply (RREP)– to first RREQ received
• Two possibilities– for bidirectional links (e.g. symmetric transmit power)
• use reverse accumulated source path in RREQ
– for asymmetric links• RREQ needed back to source if path not known
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-132
© James P.G. SterbenzITTC
24B
DSRPacket Format: RREP Packet1
• IP header [20B]– source addr [32b]
DSR target address:node replying to RREQ
– destination addr [32b]DSR originator address initiated RREQ
– TTL = [1…255] [ 8b]
• Fixed option hdr [1B]• Option
04 total length20 TOS
fragment id
TTL prot=48 header checksum
source address = target node
destination address = orig. node
frag offsetDF
MF0
nxt=59 payload lengthresvF
type=2 opt len
address[1]
address[1]
address[n]
resvL
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 67 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-133
© James P.G. SterbenzITTC
24B
DSRPacket Format: RREP Packet2
• IP header [20B]
• Fixed option hdr [1B]• Option
– type = 2 RREP [ 8b]– unique pkt id [16b]– addresses [32b]
source routecopied from RREQreturned to originator
04 total length20 TOS
fragment id
TTL prot=48 header checksum
source address = target node
destination address = orig. node
frag offsetDF
MF0
nxt=59 payload lengthresvF
type=2 opt len
address[1]
address[1]
address[n]
resvL
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-134
© James P.G. SterbenzITTC
DSROperation: Route Caching
• When source receives RREP from destination• Route cached
why?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 68 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-135
© James P.G. SterbenzITTC
DSROperation: Route Caching
• When source receives RREP from destination• Route cached
– used by all packets in a flow– can be used by other flows to same destination– all intermediate sub-paths cached for other flows
• maintained in a tree data structure
– paths cached in both directions– paths cached by other nodes overhearing RREQ and RREP
• Route caching can significantly reduce floodingDisadvantage?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-136
© James P.G. SterbenzITTC
DSROperation: Route Caching
• When source receives RREP from destination• Route cached• Route caching can significantly reduce flooding• Stale cache significantly impacts performance
– due to mobility or episodic connectivity– multiple stale routes may be tried
before success or new RREQ
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 69 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-137
© James P.G. SterbenzITTC
DSROperation: Data Transfer
• Data packets sent hop-by-hop to destination– using cached source route– each hop pops next hop from packet header– packet header length proportional to number of hops
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-138
© James P.G. SterbenzITTC
DSRAdvantages and Disadvantages
Advantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 70 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-139
© James P.G. SterbenzITTC
DSRAdvantages
• Advantages– simple algorithm– reduces flooding of control messages– paths maintained only for needed routes
• plus subpaths
– caching allows amortisation over • time: multiple flows• space: locality of flow endpoints
– good for long flows in a relatively stable network
Disadvantages?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-140
© James P.G. SterbenzITTC
DSRDisadvantages
• Disadvantages– mobility and episodic connectivity significantly impact– flooding of RREQ and RREP control messages
• may reach all nodes in a large network• collisions between adjacent nodes during flood
– insert random delays to ameliorate
– header length grows with network scale• cost paid on every packet
– intermediate node may send RREP using stale cache• poisons other caches• may be ameliorated by adding cache purge messages
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 71 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-141
© James P.G. SterbenzITTC
MANET Routing ProtocolsMR.2.4 ZRP
MR.1 Routing algorithm alternativesMR.2 Example protocols
MR.2.1 DSDVMR.2.2 AODVMR.2.3 DSRMR.2.4 ZRPMR.2.5 OLSR
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-142
© James P.G. SterbenzITTC
ZRPOverview
• ZRP (zone routing protocol)– early MANET routing protocol [Haas 1997]– not adopted by IETF MANET working group
• specification exists in expired Internet drafts
• Hybrid proactive/reactive– structured into 2-level zone based on hop-count d– proactive intra-zone over short distances ≤ d– reactive inter-zone RREQ discovery for > d hops
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 72 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-143
© James P.G. SterbenzITTC
ZRPAdvantages and Disadvantages
• Advantages– reduces overhead of RREQ for intra-zone nodes
• Disadvantages– sensitive to zone radius for given node density– high overhead due to zone overlap
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-144
© James P.G. SterbenzITTC
MANET Routing ProtocolsMR.2.5 OLSR
MR.1 Routing algorithm alternativesMR.2 Example protocols
MR.2.1 DSDVMR.2.2 AODVMR.2.3 DSRMR.2.4 ZRPMR.2.5 OLSR
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 73 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-145
© James P.G. SterbenzITTC
OLSROverview
• OLSR (optimized link state routing)– 2nd generatoin MANET routing protoco
[Clausen et al. 2001]– adopted by IETF MANET working group [RFC 3626]
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-146
© James P.G. SterbenzITTC
OLSROverview
• OLSR (optimized link state routing)– 2nd generatoin MANET routing protoco
[Clausen et al. 2001]– adopted by IETF MANET working group [RFC 3626]
• Table driven / proactive– link state
Why can’t we just use OSPF?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 74 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-147
© James P.G. SterbenzITTC
OLSRDifferences from Wired Link State
• Table driven / proactive– link state
• But there are no links in the conventional sense– OSPF and ISIS track the state of each physical link– MANET has only ephemeral links to reachable neighbours
Problem?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-148
© James P.G. SterbenzITTC
OLSRDifferences from Wired Link State
• Table driven / proactive– link state
• But there are no links in the conventional sense– OSPF and ISIS track the state of each physical link– MANET has only ephemeral links to reachable neighbours
• Problem: frequent topology & connectivity changes– result in severe overhead due to frequent flooding– every change in node pair connectivity would need LSA
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 75 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-149
© James P.G. SterbenzITTC
OLSRInformation Bases
• Each node accumulates and maintains information– repositories of information: IBs (information bases)
• OLSR IBs– multiple interface association information base– local link information base– neighborhood information base– topology information base
• Soft state that must be refreshed– time fields determine when IP tuples must be removed
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-150
© James P.G. SterbenzITTC
OLSRMultiple Interface Association IB
• Multiple interface association information base– nodes may have multiple interface addresses
• Nodes keep interface addresses from MID messages– neighbor tuple describing set of immediate neighbors– I_iface_addr interface address of node– I_main_addr main address of node– I_time expiration time of this tuple to be removed
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 76 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-151
© James P.G. SterbenzITTC
OLSRLocal Link IB
• Local link information base– neighbors discovered using HELLO messages
• Node records link tuples form incoming HELLOs– neighbor tuple describing set of immediate neighbors– L_local‐iface_addr local interface address– L_neighbor‐iface_addr neighbor interface address– L_SYM‐time time until link considered symmetric– L_ASYM‐time time until neighbor considered heard– L_time expiration time of this tuple to be removed
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-152
© James P.G. SterbenzITTC
OLSRNeighborhood IB
• Neighborhood information base– stores information about local neighbourhood– neighbor set, 2-hop neighbor set, MPR set, MPR selector set
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 77 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-153
© James P.G. SterbenzITTC
OLSRNeighborhood IB: Neighbors
• Neighborhood information base– stores information about local neighborhood
• Neighbor set
• 2-hop neighbor set• Multipoint relay (MPR) set• MPR selector set
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-154
© James P.G. SterbenzITTC
OLSRNeighborhood IB: Neighbors
• Neighborhood information base– stores information
about local neighborhood
• Neighbor set– neighbor tuple
describing set of immediate neighbors
15
21
3
6 54
7
108
91112
1314
16
17
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 78 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-155
© James P.G. SterbenzITTC
OLSRNeighborhood IB: Neighbors
• Neighborhood information base– stores information about local neighborhood
• Neighbor set– neighbor tuple describing set of immediate neighbors– N_neighbor_main_addr node main address– N_status node symmetry– N_willingness transit traffic willingness
• 2-hop neighbor set• Multipoint relay (MPR) set• MPR selector set
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-156
© James P.G. SterbenzITTC
OLSRNeighborhood IB: 2-Hop Neighbors
• Neighborhood information base– stores information about local neighborhood
• Neighbor set• 2-hop neighbor set
– 2-hop tuple describing set of 1- and 2-hop neighbors
• Multipoint relay (MPR) set• MPR selector set
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 79 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-157
© James P.G. SterbenzITTC
OLSRNeighborhood IB: 2-Hop Neighbors
• Neighborhood information base– stores information
about local neighborhood
• 2-hop neighbor set– 2-hop tuple describing
set of 1- and 2-hop neighbors
15
21
3
6 54
7
108
91112
1314
16
17
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-158
© James P.G. SterbenzITTC
OLSRNeighborhood IB: 2-Hop Neighbors
• Neighborhood information base– stores information about local neighborhood
• Neighbor set• 2-hop neighbor set
– 2-hop tuple describing set of 1- and 2-hop neighbors– N_neighbor_main_addr 1-hop node main address– N_2hop_addr 2-hop node main address– N_time expiration time of this tuple to be removed
• Multipoint relay (MPR) set• MPR selector set
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 80 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-159
© James P.G. SterbenzITTC
OLSRNeighborhood IB: MultiPoint Relays
• Neighborhood information base– stores information about local neighborhood
• Neighbor set• 2-hop neighbor set• Multipoint relay (MPR) set
– neighbors selected as MPRs– subset of neighbors that can reach all 2-hop neighbors– LSAs sent only to MPR set to reduce overhead– low-overhead MPR election algorithm heuristic
• MPR selector set
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-160
© James P.G. SterbenzITTC
OLSRNeighborhood IB: MulitPoint Relays
• Neighborhood information base– stores information
about local neighborhood
• MPR set– neighbors selected as
MPRs– subset of neighbors
that can reach all 2-hop neighbors
15
21
3
6 54
7
108
91112
1314
16
17
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 81 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-161
© James P.G. SterbenzITTC
OLSRNeighborhood IB: MPR Selector Set
• Neighborhood information base– stores information about local neighborhood
• Neighbor set• 2-hop neighbor set• Multipoint relay (MPR) set• MPR selector set
– tuple of neighbors which have selected this node as MPR– MS_neighbor_main_addr node address– MS_time expiration time of this tuple to be removed
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-162
© James P.G. SterbenzITTC
OLSRNeighborhood IB: MPR Selector Set
• Neighborhood information base– stores information
about local neighborhood
• MPR selector set– tuple of neighbors
which have selected this node as MPR
15
21
3
6 54
7
108
91112
1314
16
17
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 82 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-163
© James P.G. SterbenzITTC
OLSRTopology Information Base
• Topology information base– topology information for entire network– data structure needed by all link state routing algorithms
• Topology tuple for all nodes:– T_dest_addr node address– T_last_addr one-hop away from T_dest_addr– T_seq sequence number– T_time expiration time of this tuple to be removed
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-164
© James P.G. SterbenzITTC
12B
OLSRPacket Format and Header
• Packet consist of sequence of messages
• Packet header [1B]– length [16b]– seq# [16b]
per node interface
• Message headers [3B]• Message bodies
packet length
message header1
message1
sequence #
16B
message header2
message1
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 83 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-165
© James P.G. SterbenzITTC
OLSRPacket Format: Message Header
• Message header– type [ 8b]– Vtime [ 8b]
time to be kept valid– size [16b]– orig addr [32b]– TTL [ 8b]
limits scope of flooding– hop cnt [ 8b]– seq# [16b]
• Message body [size – 12B]
packet length
msg type
message
sequence #
16BVtime message size
TTL hop cnt msg sequence #
originator address
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-166
© James P.G. SterbenzITTC
OLSRMessage Types
• OLSR packets are for control only– as for OSPF– may be encapsulated in another L3 protocol
• UDP port 698 reserved for OLSR• typically encapsulated in UDP in IP in MAC frame
• Message types (OLSR is only a routing protocol)1 = HELLO neighbor discovery and keepalive2 = TC topology control: link state advertisements3 = MID multiple interface declaration (for a given node)4 = HNA host and network association (OLSR–external GW)
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 84 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-167
© James P.G. SterbenzITTC
OLSRMessage Format: HELLO
• HELLO message– HTime [ 8b]
HELLO time interval– Willing [ 8b]
willingness to forward[0=never … 7=always]
– link code [ 8b]type (symmetry)
– link msg size[1 + #(intfc addr)B]
– neigh intfc addr [32b]list of addresses
packet length
type = 01
sequence #
16BVtime message size
TTL hop cnt msg sequence #
originator address
reserved WillingHTime
link msg sizereservedlink code
neighbor interface address
neighbor interface address
link msg sizereservedlink code
neighbor interface address
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-168
© James P.G. SterbenzITTC
OLSRMessage Format: TC
• HELLO message– HTime [ 8b]
HELLO time interval– Willing [ 8b]
willingness to forward[0=never … 7=always]
– link code [ 8b]type (symmetry)
– link msg size[1 + #(intfc addr)B]
– neigh intfc addr [32b]list of addresses
packet length
type = 01
sequence #
16BVtime message size
TTL hop cnt msg sequence #
originator address
reserved WillingHTime
link msg sizereservedlink code
neighbor interface address
neighbor interface address
link msg sizereservedlink code
neighbor interface address
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 85 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-169
© James P.G. SterbenzITTC
OLSRPacket Processing
• Node examines Packet Type of received packet– duplicate tuple record created to ignore subsequent dups
1. Packets with invalid short length discarded2. Packets with TTL ≤ 0 discarded3. Duplicate packets discarded; else Type decoded4. Message processed, and forwarded if applicable
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-170
© James P.G. SterbenzITTC
OLSRPacket Forwarding
1. If sender address not 1-hop neighbor, drop2. If duplicate, check to see if retransmission3. Drop if duplicate not retransmitted4. If from MPR selector and TTL > 1 forward5. Update duplicate tuple record6. Decrease TTL by 17. Increase hop count by 18. Broadcast on all interfaces
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 86 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-171
© James P.G. SterbenzITTC
OLSRMessage Format: Topology Control
• TC message– ANSN [ 8b]
advertized neighbor sequence number: incremented with every topology change so stale updates don’t affect net
– neigh main addr [32b]list of MPR neighbors of originating node
packet length
type = 02
sequence #
16BVtime message size
TTL=255 hop cnt msg sequence #
originator address
ANSN
advertized neighbor main address
advertized neighbor main address
reserved
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-172
© James P.G. SterbenzITTC
OLSRAdvantages and Disadvantages
Advantages?
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 87 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-173
© James P.G. SterbenzITTC
OLSRAdvantages and Disadvantages
• Advantages+ lower overhead than other proactive algorithms+ scalable to large networks
• hierarchy could be added for very large networks
Disadvantages?
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-174
© James P.G. SterbenzITTC
OLSRAdvantages and Disadvantages
• Advantages+ lower overhead than other proactive algorithms+ scalable to large networks
• hierarchy could be added for very large networks
• Disadvantages– maintains routes even for unneeded paths– link state only up or down; no notion of weak connectivity– overhead of periodic LSAs
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 88 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-175
© James P.G. SterbenzITTC
OLSRDeployment
• Used in some mesh deployments– OpenWRT on Linksys WRT54GS– but this is targetted for fixed mesh networking
• rather than MANET
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-176
© James P.G. SterbenzITTC
MANET Routing Further Reading
• Charles E. Perkins, ed.,Ad Hoc Networking, 2nd ed.,Pearson Addison Wesley, Boston, 2001
• Martha Steenstrup, ed.,Routing in Communications Networks,Prentice-Hall, 1995
KU EECS 882 – Mobile Wireless Networking – MANET Routing Algorithms and Protocols
– 89 –
29 October 2009 KU EECS 882 – Mobile Wireless Nets – MANET Routing MWN-MR-177
© James P.G. SterbenzITTC
MANET RoutingAcknowledgements
Some material in these foils is based on the textbook• Murthy and Manoj,
Ad Hoc Wireless Networks:Architectures and Protocols
Some information in these foils derived from the tutorial• Nitin H. Vaidya
Mobile Ad Hoc Networks: Routing, MAC, and Transport Issuesavailable fromhttp://www.crhc.uiuc.edu/wireless/talks/2006.Infocom.ppt
Some material in these foils enhanced from EECS 780 foils