1. Ubiquitous Networks Mobile Ad Hoc Networks
Lynn Choi
Korea University

2. Mobile Ad hoc Network Ad hoc network consisting of mobile nodes
No infrastructure such as base stations or access points
Self-organizing networks
Mobile nodes
Any node can be mobile anytime
Each node is assumed to be more powerful compared to sensor nodes in terms of computation power, memory, and energy source
Examples – automobile networks, mobile user networks
Peer-to-peer communication patterns
Any node can send data packet to any other node
Neither sources nor sinks
Networking issues
Not easy to use topology-based routing since the mobility of nodes may change the topology anytime. Node mobility may lead to frequent routing failures.

3. Routing Protocols for MANET
Two main groups of routing protocols
Proactive protocols (Table driven)
Destination-Sequenced Distance Vector (DSDV)
Optimized Link State Protocol (OLSR) [RFC 3626]
Topology Broadcast Based on Reverse-Path Forwarding (TBRPF) [RFC 3684]
Reactive protocols (On-demand)
Dynamic Source Routing (DSR)
Ad-hoc On-demand Distance Vector (AODV) [RFC 3561]
Hybrid Protocols
Zone Routing Protocols (ZRP)

4. Classification of MANET Routing Protocols
Unicast-Routing Protocol for MANET (Topology-based)
Table-Driven/
Proactive
Hybrid
On-Demand-driven/Reactive
Clusterbased/
Hierarchical
Distance-
Vector
Link-
State
ZRP
DSR
AODV
TORA
LANMAR
CEDAR
DSDV
OLSR
TBRPF
FSR
STAR
MANET: Mobile Ad hoc Network
(IETF working group)
Source: IETF

5. Routing Protocols for Ad-hoc Networks
Proactive protocols
Create and maintain routes to all hosts proactively
Conventional approach to routing protocols
Table-driven
Distance vector and link state protocols send periodic routing advertisement
(+) Lower latency – since routes are maintained at all times
(-) Higher overhead – due to continuous route updates
(-) Convergence problems
Reactive Protocols
Obtain a route only when needed
Source-driven (on-demand)
Eliminate periodic updates – saves power and bandwidth
Potentially increased latency while routes found or repaired
(+) Low overhead – routes are determined on demand
Skipping periodic updates saves energy
(+) Quicker response to topology changes
(-) Higher latency – need route discovery
Which approach achieves a better trade-off depends on the traffic and mobility patterns
In general, reactive protocols are more popular in MANET

6. Distance-Vector Routing
Two classes of proactive routing algorithms
Link state routing
Each node maintains a global view of the network topology with a cost for each link
Shortest-path computation
Each node periodically broadcasts link costs to its outgoing links
Distance vector routing
Every node maintains the distance and the direction to each destination
No global view of the topology
Each node periodically sends routing table to all neighbors
Each node maintains a routing table containing
List of all available destinations
Distance to each each destination
Next hop to reach a destination
The succession of next hops leads to a destination
Each node periodically broadcasts its current estimate of the shortest distance to each available destination to all of its neighbors
Typical representative
Distributed Bellman-Ford (DBF)
Computes single-source shortest paths in a weighted directed graph

7. Bellman-Ford Algorithm
Initialize Single Source (G,s)
for each vertex v V[G]
do d[v] 
[v]  NIL
d[s]  0
Bellmann-Ford (G,w,s)
Initialize-Single-Source (G,s)
for i  1 to |V[G]|-1
do for each edge (u,v) E[G]
do Relax (u,v,w)
for each edge (u,v) E[G]
do if d[v] > d[u]+w(u,v)
then return FALSE
return TRUE
Relax (u,v,w)
if d[v] > d[u] + w(u,v)
then d[v]  d[u] + w(u,v)
[v]  u
d[v]: distance
[v]: next hop
Check for negative weight cycles

8. Bellman-Ford Algorithm

9. Bellman-Ford Routing Advantages Routing loop problem
Computationally efficient
Easy to implement
Routing loop problem
Can cause loops
At time t0, B detects link cost changes (4 to 60) and computes its new minimum distance to A
DB(A) = min { 60, 1 + DC(A)} = min{60, 1+5} = 6
Without the gloval view of the network, node B concludes that it can reach A in 6 hops!
So, in order to get to A, B would route through C and C routes through B, causing a routing loop
Count-to-infintiy problem
At time t1, B informs C of its new distance vector
At time t2, C receives B‘s new distance vector and computes its new minimum distance to A
DC(A) = min { 50, 1 + DB(A)} = min{60, 1+6} = 7
At time t3, C informs B of its new distance vector
At time t4, B receives C‘s new distance vector and computes its new minimum distance to A, which is 8
This continues 60 B 4 1 C A 50

10. DSDV Design goals: Approach: Keep the simplicity of Bellman-Ford
Address the routing loop and count-to-infinity problems
Approach:
Guarantee loop freeness
New table entry for Destination Sequence Number
Allow fast reaction to topology changes
Make immediate route advertisement on significant changes in routing table
But wait for advertising unstable routes (damping fluctuations)

11. DSDV Routing Table Sequence number Install Time Stable Data
Destination
Next
Metric
Seq. Nr
Install Time
Stable Data A
A-550
001000
Ptr_A B 1
B-102
001200
Ptr_B C 3
C-588
Ptr_C D 4
D-312
Ptr_D
Sequence number
Originated from destination.
Ensures loop freeness.
Install Time
When entry was made (used to delete stale entries from table)
Stable Data
Pointer to a table holding information on how stable a route is. Used to damp fluctuations in network.

12. Route Advertisements & Updates
Advertise to each neighbor its own routing information
Destination Address
Metric = Number of Hops to Destination
Destination Sequence Number
Rules to set sequence number information
On each advertisement increase own destination sequence number (use only even numbers)
If a node is no more reachable (timeout) increase sequence number of this node by 1 (odd sequence number) and set metric = 
Immediate versus periodic updates
Immediate advertisements
Information on new Routes, broken Links, metric change is immediately propagated to neighbors.
Otherwise, periodic updates
Full/Incremental Update:
Full Update: Send all routing information from its own table.
Incremental Update: Send only entries that has changed. (Make it fit into one single packet)

13. Route Selection Route selection Stale entries
Update information is compared to its own routing table
Select route with higher destination sequence number.
This ensure to use always newest information from destination)
Select the route with better metric when sequence numbers are equal.
Stale entries
Defined to be entries that have not been updated the last few update periods
Stale entries are deleted at the same time when routing updates are applied to the routing table
Any route using that host as a next hop is deleted, including the route indicating that host as the actual destination

14. Receiving Fluctuating Routes
What might happen:
MH9 broadcasts update information to MH Collections I and II
MH2 transmits new routing information to MH4
MH4: new sequence number  routing table update  broadcast update
MH6 transmits new routing information to MH4, same sequence number, better metric
MH4: same seq.no., better metric  update routing table  broadcast update
MH9
Mobile Host
Collection I
Mobile Host
Collection II
MH2
MH6
MH4

15. Damping Fluctuation Causes for Fluctuation: Solution
Many hosts with irregular updates
Different propagation speed
Different transmission intervals
Broadcasts are asynchronous events
Solution
Keep a route settling time table in each node with a time to wait for a route with a better metric before advertising the update message.
Settling time
Calculated by maintaining a running weighted average over the most recent updates of the routes for each destination.

16. Example of DSDV in operation
MH3
MH4
MH5
MH2
MH8
MH6
MH7
MH1
Destination
Next Hop
Metric
Seq. No
MH4
S406_MH4
MH1
MH2 2
S128_MH1 1
S564_MH2
MH3
S710_MH3
MH5
MH6
S392_MH5
S076_MH6
MH7
S128_MH7
MH8 3
S050_MH8
MH4 advertised table:

17. Example of DSDV in operation
MH3
MH4
MH5
MH2
MH8
MH6
MH7
MH1
MH1
Update triggered by MH1 , broadcasted to MH7 and MH8
On detection of broken link: Immediate incremental update triggered by MH2 with odd sequence number and infinite metric
Updates are propagated through the network

18. Example of DSDV in operation
MH3
MH4
MH5
MH2
MH8
MH6
MH7
MH1
Destination
Next Hop
Metric
Seq. No
MH4
S516_MH4
MH1
MH6 3
S238_MH1
MH2 1
S674_MH2
MH3 2
S820_MH3
MH5
S502_MH5
S186_MH6
MH7
S238_MH7
MH8
S160_MH8
MH4 advertised table (updated):

19. Simulation results Simulation by Broch, Maltz, Johnson, Hu, Jetcheva
DSDV fails to converge if nodes don‘t pause for at least 300 seconds
Packet delivery ratio is in the range of 70%-92% at higher rate of mobility
Packet loss is mainly caused by stale routing entries
Routing overhead is approximately constant, regardless of movement rate or traffic load
Nearly optimal path can be selected in routing procedure

20. Stability and Scalability
DSDV requires a full dump update periodically
 DSDV is not efficient in route updates
DSDV limits the number of nodes that can join the network
Whenever topology of a network changes, DSDV is unstable until update packets propagate through the network
Conclusion
DSDV is effective for creating ad-hoc networks for small populations of mobile nodes
DSDV is a fairly brute force approach, because connectivity information needs periodical update througout the entire network
Current Status
DSDV is a well-known routing algorithm for ad hoc network routing
No standard specifications or commercial implementations available
Many improved protocols based on DSDV have been developed
Example: AODV: Ad-hoc On-Demand Distance Vector Routing

21. Optimized Link State Routing (OLSR) Protocol
[RFC 3626]
The overhead of flooding link state information is reduced by requiring fewer nodes to forward the information
Conventional link-state routing
For each node, disseminate its links to all other nodes
Use SPF algorithm to generate routing table
Optimized variant of conventional link state routing
A broadcast from node X is only forwarded by some of its links (multipoint relays)
Modified Shortest Path First routing

22. Topology Broadcast Based on Reverse Path Forwarding
[RFC 3626]
A variation of link state algorithm
Idea
The communication cost can be reduced if the updates are sent along spanning trees despite the additional cost for maintaining these trees.
Most link state algorithms are based on flooding.
Transmits only the differences between the previous network state and the current network state
Reference
B. Bellur, and R.G. Ogier "A Reliable, Efficient Topology Broadcast Protocol for Dynamic Networks," Proc. IEEE INFOCOMM ’99, pp. 178–186.

23. Dynamic Source Routing (DSR) - Route Discovery
When node S wants to send a packet to node D, but does not know a route to D, node S initiates a route discovery
Source node S floods Route Request (RREQ)
RREQ contains initiator, target, and unique request ID determined by RREQ initiator
RREQ also contains a record listing address of each intermediate node through which RREQ has been forwarded
When a node receives RREQ, if it is the target, it sends a Route Reply (RREP) message
Otherwise, if the node receiving RREQ recently saw another RREQ from the initiator having the same request ID, or if it finds that its own address is already listed in the route record, it discard the RREQ
If not, the node appends its own address to the route record in the RREQ and forward it

24. Route Discovery in DSR
Building of the route record during route discovery N2
N1-N2
Destination
N1-N2-N5 N8 N5 N1
Source
N1-N3-N4
N1-N3-N4-N7 N1 N7
N1-N3-N4 N4
N1-N3 N1
N1-N3-N4-N6 N3
N1-N3-N4 N6

25. Route Reply in DSR
Propagation of the route reply with the route record N2
N1-N2-N5-N8
Destination
N1-N2-N5-N8 N8 N5
N1-N2-N5-N8
Source N1 N7 N4 N3 N6

26. Dynamic Source Routing (DSR) - Data Delivery
RREP can be sent by reversing the route in RREQ
Assume that links are bi-directional
If unidirectional links are allowed, then RREP may need a route discovery for S from node D, unless D already knows a route to S (in its route cache)
RREP includes the route from S to D
The entire route is included in the packet header (hence source routing)
Intermediate nodes use the source route included in the packet to determine to whom a packet should be forwarded
On receiving RREP, intermediate nodes cache the route included in RREP
Route Maintenance - Route Error packet
A node generates Route Error packet, when the data link layer encounters a fatal transmission problem
Receiving node removes
The hop in error from the node’s route cache
All routes containing the hop

27. Data Delivery in DSR Packet header size grows with route length N2 N8
DATA N1-N2-N5 N2
Destination
DATA N1-N2-N5-N8
DATA N1-N2 N8 N5
Source N1 N7 N4 N3 N6

28. Ad Hoc On-Demand Distance Vector Routing (AODV)
<RFC3561>
DSR includes source routes in packet headers
Resulting large headers can sometimes degrade performance
Particularly when the data content of a packet is small
Also, packet header size grows with route length
AODV attempts to improve on DSR by maintaining routing tables at the nodes,
So that data packets do not have to contain routes
When node S want to send a packet to node D, it checks its route table to determine whether it has a route to D
If S has a route to D, S forwards packets to the next hop toward D
If S does not have route to D, S initiates route discovery

29. AODV Route Discovery
Source node S floods Route Request (RREQ), which contains
Source IP address
Destination IP address
Source node’s current sequence number
Last known sequence number from the destination
Broadcast ID
Hop count
After sending RREQ, S sets a timer to wait for a reply
When a node receives RREQ
It checks whether it has already seen the RREQ by noting source address/broadcast ID pair. If so, it discard it
Otherwise, it setup a reverse path pointing towards the source
AODV assumes bi-directional links
Sets up reverse route entry which contains source address, sequence #, # of hops to S, neighbor address from which the RREQ was received
If route entry is not used within specified lifetime, the route entry is deleted to prevent stale routing
Timeout should be long enough to allow RREP to come back

30. AODV - Route Discovery A node can send a RREP if
1) it has an unexpired entry for the destination in its route table, and
2)it knows a more recent path than the one previously known to sender S
To determine whether the path known to an intermediate node is more recent, destination sequence number is used
A node which knows a route, but with a smaller sequence number, cannot send RREP
If the node cannot send RREP
It increments RREQ’s hopcount and then broadcast the RREQ to its neighbors

31. Route Requests in AODV Broadcast transmission S E F B C J A G H D K I
Represents transmission of RREQ
Represents links on Reverse Path

32. AODV - Forward Path Setup
When the destination receives a RREQ,
It replies by sending a RREP, where it places the current sequence #, initializes hop count to 0, and places the length of time the route is valid in lifetime field
When an intermediate node receives RREP,
It sets up a forward path entry to destination in its route table, which contains destination address, neighbor address from which the RREP arrived, hop-count (distance) to the destination, and lifetime
To obtain the distance to destination, each node increments the value in hop-count by one
Lifetime in route table is set to the lifetime contained in the RREP
Each time the route is used, its lifetime is updated
If the route is not used within the specified lifetime, it is deleted
A node may receive RREP for a given destination from more than one neighbor
It forwards first RREP and forward a later RREP only if the RREP contains a larger destination sequence number or a smaller hop-count
RREP travels along the reverse path setup when RREQ is forwarded
S can begin data transmission as soon as the first RREP is received and later updates its routing info if it discovers a better route

33. Route Reply in AODV S E F B C J A G H D K I
RREP S E
RREP F B
RREP C J
RREP A G H D K I
Represents links on path taken by RREP

34. Data Delivery in AODV DATA DATA S E DATA F B C J DATA A G H D K I
Routing table entries used to forward data packet.
Route is not included in packet header.
At most one next-hop per destination maintained at each node

35. Zone Routing Protocol (ZRP)
All nodes have their own Zone, based on Zone Radius.
Intra Zone : Proactive
Inter Zone : Reactive L K s M C B s G D E A H F I
Intra Zone
Inter Zone