Adhoc routing protocols

Routing Protocols for Ad-Hoc Networks
Ad-hoc On-Demand Distance Vector Routing
&
DSR: The Dynamic Source Routing Protocol
for Multi-Hop Wireless Ad Hoc Networks
November 2011,
Giorgos Papadakis & Manolis Surligas

Outline
 Ad-Hoc networks
 Ad-hoc routing algorithms
 Ad-Hoc on-demand Distance Vector Routing (AODV)
 Dynamic Source Routing (DSR)
 Comparison of AODV and DSR

Ad Hoc Networks
Wireless networks can be divided in two fundamental
categories:
 Infrastructure-based
Wireless clients connecting to a base-station (APs,
Cell Towers) that provides all the traditional network
services (routing, address assignment)
 Infrastructure-less
The clients themselves must provide all the traditional
services to each other

Ad Hoc Networks
Ad-hoc networks main features:
 Decentralized
 Do not rely on preexisting infrastructure
 Each node participates in routing by
forwarding data to neighbor nodes
 Fast network topology changes due to nodes’
movement

Ad Hoc Networks
Why do we need ad-hoc networks?
 More laptop users
 More smartphones users (e.g.. Android phones,
iPhones)
 More devices with Wi-Fi-support (e.g.. televisions, hi-
fi, home-theaters, media servers etc.)
 Moving users, vehicles, etc.
 Outdoors places
 In all these occasions there is no centralized
infrastructure (such APs)
 So ad-hoc network is a necessity

Ad Hoc Networks
An Ad-hoc networkAn infrastructure wireless network

Ad-hoc routing algorithms
Hottest routing algorithm categories:
 Pro-active (table-driven) routing
Maintains fresh lists of destinations & their routes by periodically
distributing routing tables
Disadvantages:
1. Respective amount of data for maintenance
2. Slow reaction on restructuring and failures
(e.g. OSLR, DSDV)
 Reactive (on-demand) routing
On demand route discovery by flooding the network with Route
Request packets
Disadvantages:
1. High latency time in route finding
2. Flooding can lead to network clogging
(e.g. AODV, DSR)

Ad-hoc routing algorithms
Discuss and comparison
1. Ad-Hoc on-demand Distance Vector Routing (AODV)
2. Dynamic Source Routing (DSR)

Outline
 Ad-Hoc networks
 General info
 Path Discovery
 Path Maintenance
 Local Connectivity Maintenance
 Conclusion

(AODV) General info
 Reactive algorithms like AODV create routes on-
demand. They must however, reduce as much as
possible the acquisition time
 We could largely eliminate the need of periodically
system-wide broadcasts
 AODV uses symmetric links between neighboring
nodes. It does not attempt to follow paths
between nodes when one of the nodes can not hear
the other one

(AODV) General info
 Nodes that have not participate yet in any packet
exchange (inactive nodes), they do not maintain
routing information
 They do not participate in any periodic routing table
exchanges

(AODV) General info
 Each node can become aware of other nodes in its
neighborhood by using local broadcasts known as
hello messages
 neighbor routing tables organized to :
1. optimize response time to local movements
2. provide quick response time for new routes
requests

(AODV) General info
AODV main features:
 Broadcast route discovery mechanism
 Bandwidth efficiently (small header information)
 Responsive to changes in network topology
 Loop free routing

(AODV) Path Discovery
 Initiated when a source node needs to
communicate with another node for which it has no
routing info
 Every node maintains two counters:
 node_sequence_number
 broadcast_id
 The source node broadcast to the neighbors a
route request packet (called RREQ)

 RREQ structure
<src_addr, src_sequence_#, broadcast_id, dest_addr,
dest_sequence_#, hop_cnt>
 src_addr and broadcast_id uniquely identifies a RREQ
 broadcast_id is incremented whenever source node
issues a RREQ
 Each neighbor either satisfy the RREQ, by sending
back a routing reply (RREP), or rebroadcast the RREQ
to its own neighbors after increasing the hop_count by
one.

 If a node receives a RREQ that has the same
<src_addr, broadcast_id> with a previous RREQ it
drops it immediately
 If a node cannot satisfy the RREQ, stores:
 Destination IP
 Source IP
 broadcast_id
 Expiration time (used for reverse path process)
 src_sequence_#

1. Reverse Path Setup
 In each RREQ there are:
 src_sequence_#
 the last dest_sequence_#
 src_sequence_# used to maintain freshness
information about the reverse route to the source
 dest_sequnece_# indicates how fresh a route must
be, before it can be accepted by the source

1.Reverse Path Setup (continue)
 As RREQ travels from source to many destinations, it
automatically sets up the reverse path, from all nodes
back to the source.
But how does it work?
 Each node records the address of the neighbor from which it
received the first copy of the RREQ
 These entries are maintained for at least enough time, for the
RREQ to traverse the network and produce a reply

1.Reverse Path Setup (continue)
U
D
Z
Y
W
S
V
S
D
Z
W
ZW
Source node
Destination node
Neighbor nodes
S sends RREQ
Figure 1
W, Y can not satisfy RREQ
i. Set up reverse path
ii. Rebroadcast RREQ to
neighbors
Z, V, U can not satisfy RREQ
i. Set up reverse path
ii. Rebroadcast RREQ to
neighbors
RREQ reached destination
Reversed path is fully set up
From which RREP can travel
back to S

2. Forward Path Setup
 A node receiving a RREP propagates the first RREP
for a given source towards the source (using the
reverse path that has already established)
 Nodes that are not in the path determined by the
RREP will time out after 3000 ms and will delete the
reverse pointers

2. Forward Path Setup (continue)
U
D
Z
Y
W
S
V
S
D
Z
W
WZ
Source node
Destination node
Z has a reversed path to W
Figure 2
ZW W has a forward path to Z
D replies with a
RREP to Z
Z receives RREP
and set up a
forward pointer
The same
for the
other
nodes
Time out

2. Forward Path Setup (Conclusion)
 Minimum number of RREPs towards source
 The source can begin data transmission as soon as
the first RREP received and update later its routing
information if it learns of a better route

(AODV) Path Maintenance
 Movement of nodes not lying along an active path does NOT
affect the route to that path's destination
 If the source node moves, it can simply re-initiate the route
discovery procedure
 If the destination or some intermediate node moves, a
special RREP is sent to the affected nodes
 To find out nodes movements periodic hello messages can be
used, or (LLACKS) link-layer acknowledgments (far less
latency)

(AODV) Path Maintenance
 When a node is unreachable the special RREP that
is sent back towards the source, contains a new
sequence number and hop count of ∞
U
D
Z
Y
S
V
Z
W
Figure 3
Link between Z
and D fails
Z sents a
special RREP
So do W
So now source must find a new path. To do that, it sents a RREQ with a new greater
sequence number

(AODV) Local Connectivity
Maintenance
 Nodes learn of their neighbors in one or two ways:
1. Whenever a node receives a broadcast from a
neighbor it update its local connectivity
information about this neighbor
2. If a neighbor has not sent any packets within
hello_interval it broadcasts a hello message,
containing its identity and its sequence number

Maintenance
How hello messages work:
 Hello messages do not broadcasted outside the
neighborhood because the contain a TTL (time to
leave) value of 1.
 Neighbors that receive the hello message update
their local connectivity information to the node that
have broadcasted the hello message

Maintenance
How hello messages work: (continue)
 Receiving a hello from a new neighbor, or failing to
receive allowed_hello_loss (typically 2) consecutive
hello messages from a node previously in the
neighborhood, indicates that the local connectivity
has changed

(AODV) Conclusion
AODV main features:
 Nodes store only the routes they need
 Need for broadcast is minimized
 Reduces memory requirements and needless
duplications
 Quick response to link breakage in active routes
 Loop-free routes maintained by use of destination
sequence numbers
 Scalable to large populations of nodes

Outline
 Ad-Hoc networks
 General
 Basic Route Discovery
 Basic Route Maintenance
 Conclusion

(DSR) General
Two main mechanisms that work together to allow the
discovery and maintainance of source routes:
 Route discovery
 Route maintainance

(DSR) General
Route discovery:
 Is the mechanism by which a source node S, obtains
a route to a destination D
 Used only when S attempt to send a packet to D and
does not already knows a route to D

(DSR) General
Route maintainance:
 Is the mechanism by which source node S is able to
detect if the network topology has changed and can
no longer use its route to D
 If S knows another route to D, use it
 Else invoke route discovery process again to find a
new route
 Used only when S wants to send a packet to D

(DSR) General
 Each mechanism operate entirely on demand
 DSR requires no periodic packets of any kind at any
level
 Uni-directional and asymmetric routes support
(e.g. send a packet to a node D through a route and receive a
packet D from another route)

(DSR) Basic Route Discovery
When S wants to sent a packet to D:
 it places in the header of the packet a source route
giving the sequence of hops that the packet should
follow on its way to D
 S obtains a suitable source route by searching its route
table
 If no route found for D, S initiate the Route Discovery
protocol to dynamically find a new route to D

Sender
 Broadcasts a Route Request Packet (RREQ)
 RREQ contains a unique Request ID and the address of the
sender
Receiver
 If this node is the destination node, or has route to the
destination send a Route Reply packet (RREP)
 Else if is the source, drop the packet
 Else if is already in the RREQ's route table,
drop the packet
 Else append the node address in the RREQ's route table
and broadcast the updated RREQ

U
D
Z
Y
W
S
V
S
D
Z
W
ZW
Source node
Destination node
Neighbor nodes
S sends RREQ
Figure 4
RREQ packet
Id=2, {S}
Id=2, {S}
Id=2, {S, W}
Id=2, {S, Y}
Id=2, {S, Y}
Id=2, {S, W, Z}

When a RREQ reaches the destination node, a RREP
must be sent back to source
The destination node:
 Examine its own Route Cache for a route back to source
 If found, it use this route to send back the RREP
 Else, the destination node starts a new Route Discovery
process to find a route towards source node
 In protocols that require bi-directional links like 802.11, the
reversed route list of the RREQ packet can be used, in order to
avoid the second Route Discovery

(DSR) Basic Route Maintenance
Each node transmitting a packet:
 is responsible for confirming that the packet has been received
by the next hop along the source route
 The confirmation it is done with a standard part of MAC layer
(e.g. Link-level ACKs in 802.11)
 If none exists, a DSR-specific software takes the
responsibility to sent back an ACK
 When retransmissions of a packet in a node reach a maximum
number, a Route Error Packet (RERR) is sent from the node back
to the source, identifying the broken link

The source:
 Removes from the routing table the broken route
 Retransmission of the original packet is a function of
upper layers (e.g. TCP)
 It searches the routing table for another route, or start
a new Route Discovery process

U
D
Z
Y
W
S
V
S
D
Z
W
ZW
Source node
Destination node
Neighbor nodes
Figure 5
RERR packet
Link fails
Intermediate
node sents a
RERR
RERR(Z, D)
RERR(Z, D)
Route Table
D: S, W, Z, D
V: S, Y, V

(DSR) Conclusion
 Excellent performance for routing in multi-hop wireless
ad hoc networks
 Very low routing overhead even with continuous rapid
motion, which scales to :
1. zero when nodes are stationary
2. the affected routes when nodes are moving
 Completely self-organized & self-configuring network
 Entirely on-demand operation. No periodic activity of any
kind at any level

Comparison of AODV and DSR
Main common features:
 On-demand route requesting
 Route discovery based on requesting and replying
control packets
 Broadcast route discovery mechanism

Main common features: (continue)
 Route information is stored in all intermediate nodes
along the established path
 Inform source node for a broken links
 Loop-free routing

Main differences:
 DSR can handle uni and bi-directional links, AODV uses
only bi-directional
 In DSR, using a single RREQ - RREP cycle, source and
intermediate nodes can learn routes to other nodes on
the route
 DSR maintains many alternate routes to the destination,
instead of AODV that maintains at most one entry per
destination

Main differences: (continue)
 DSR doesn’t contain any explicit mechanism to expire
stale routes in the cache , In AODV if a routing table
entry is not recently used , the entry is expired
 DSR can’t prefer “fresher” routes when faced multiple
choices for routes. In contrast, AODV always choose
the fresher route (based on destination sequence
numbers)

Main differences: (continue)
 DSR’s RREQ has variable length depending on the nodes
that the packet has traveled. AODV’s RREQ size is
constant
 As a result DSR’s header overhead may increase as more
nodes become active, so we expect that AODV
throughput in those scenarios to be better

Test bench set up:
 100 nodes, some of them as sources
 Nominal bit rate of 2 Mb/s
 Nominal node range of 250 m
 Continuously moving nodes

Application and routing statistics for an example scenario for a network of 100
nodes with continuous mobility and 40 sources
Performance
metrics
DSR AODV
Packets delivered
/Packets sent (%)
56.88 83.66
Average delay (s) 1.36 0.26
Routing Packets DSR AODV
Route requests 37774 228094
Route replies 82710 17753
Route errors 26591 9808
Total 147075 255655

Conclusion
 DSR outperforms AODV in less stressful situations
(i.e., smaller number of nodes and lower load and/or
mobility)
 AODV outperforms DSR in more stressful situations
(e.g., more load, higher mobility)
 DSR commonly generates less routing load than AODV
 Poor delay and throughput of DSR due to lack of any
mechanism to expire stale routes or determine the
freshness of routes

Adhoc routing protocols

More Related Content

What's hot (20)

Similar to Adhoc routing protocols (20)

Recently uploaded (20)

Adhoc routing protocols