Routing

Pritee Parwekar, Jaypee Institute of Information Technology

 Routing Techniques
• Sharing information between physically separated hosts/
sources requires both physical connections between these
hosts in terms of cables, links, etc. and a common language,
called protocol, to make these hosts understand each other.
• Although WSNs have some similarities with traditional
networks, currently available routing protocols can not be
directly applied to WSNs because of some characteristics of
WSNs listed below:

Pritee Parwekar, Jaypee Institute of

2
Information Technology
 Wireless Sensor Networks
• low processing capacity
• difficult operations conditions
• limited energy source
• low bandwidth
• huge population in WSNs
• Non-predetermined position of sensor nodes

Pritee Parwekar, Jaypee Institute of

3
Information Technology
 Design Issues of WSNs

• Design of routing protocols in WSNs is influenced by many

challenging factors to be addressed. Some of them are:
– Node deployment
– Data reporting method
– Node/link heterogeneity
– Fault tolerance
– Scalability
– Transmission media
– Connectivity
– Data aggregation
– Quality of service

Pritee Parwekar, Jaypee Institute of

4
Information Technology
Pritee Parwekar, Jaypee Institute of
5
Information Technology
 Routing methods depending on networks
structure:
• Flat routing
• Hierarchical routing
• Location-based routing

Pritee Parwekar, Jaypee Institute of

6
Information Technology
Routing depending on protocol operation.

• Negotiation-based routing
• Multipath-based routing
• Query-based routing
• QoS-based routing
• Coherent based routing

Pritee Parwekar, Jaypee Institute of

7
Information Technology
 In addition to above, depending on route
determination, routing protocols cans be classified
into four categories

• Proactive routing protocols: All routes are computed

before they are used.
• Reactive routing protocols: Routes are computed as
they are needed.
• Hybrid routing protocols: uses both proactive and
reactive routing protocols.
• Cooperative routing protocols: Nodes send data to a
central node where more processing power and
route information is available.
Pritee Parwekar, Jaypee Institute of
8
Information Technology
 Flat Routing:Flooding and Gossiping
Flooding :

• Node A sends data to all its neighbors

• Neighbors of A sends data to all their
neighbors
• Until all nodes receive data

Pritee Parwekar, Jaypee Institute of

9
Information Technology
 Flooding
• Each node acts as both a transmitter and a
receiver.
• Each node tries to forward every message to every
one of its neighbors except the source node.
• Disadvantage : It is very wasteful in terms of the
networks total bandwidth.
• Advantage: The main advantage of flooding is the
increased reliability provided by this routing
method.

Pritee Parwekar, Jaypee Institute of

10
Information Technology
 Flat Routing:Flooding and Gossiping
Gossiping :

• Bases on flooding
• Data is only forwarded to one randomly
selected neighbor
• Saves energy

Pritee Parwekar, Jaypee Institute of

11
Information Technology
Gossiping Example

Pritee Parwekar, Jaypee Institute of

12
Information Technology
• Flat Routing Protocols
– Directed Diffusion
– SPIN
• Hierarchical Routing Protocols
– LEACH
– PEGASIS
– TEEN

13
Address and Data-centric

14
 Sensor Protocols for Information via
Negotiation ( SPIN )

• This protocol uses the idea of distributing only

the data that other nodes do not have,
assuming the nodes in close proximity have
similar data.
• Protocol starts when SPIN node gathers new
data.
• It uses ADV ,REQ and DATA msg (or signals)

15
 SPIN-1 - three types of messages
• ADV
– When a SPIN node has data to share, it can advertise an
ADV message containing meta-data.
• REQ
– A SPIN node sends an REQ message when it wishes to
receive some actual data.
• DATA
– DATA messages contain actual sensor data with a meta-
data header.

Pritee Parwekar,1J7aypee Institute

of
Information Technology
 The SPIN-1 Protocol
• Steps

C
RDAE
ADQ
ADV DARA
D B A
TV EDTQA
V
ADV
E ARD
EDA
QTV
F
A
 The SPIN-2 Protocol
• When energy is plentiful, SPIN-2 nodes
communicate using the same 3-stage protocol
as SPIN-1 nodes.

• When a SPIN-2 node observes that its energy

is approaching a low-energy threshold, it
adapts by reducing its participation in the
protocol.

Pritee Parwekar,1J9aypee Institute

of
Information Technology
Pritee Parwekar, Jaypee Institute of
19
Information Technology
 Advantage
• Important advantage of SPIN protocol is that each
node only knows its single-hop neighbors therefore
topological changes in network localized, i.e. does
not affect whole network. On the other hand, SPIN
protocol does not guarantee delivery of data because
intermediate nodes between source and destination
nodes may not be interested in advertised data,
therefore such data may not be forwarded to
destination.

Pritee Parwekar, Jaypee Institute of

20
Information Technology
 SPIN Protocol
• Note.Wendi Heinzelman, Joanna Kulik, and
Hari Balakrishnan, Adaptive Protocols for
Information Dissemination in Wireless Sensor
Networks, Proc. 5th ACM/IEEE Mobicom
Conference, Seattle, WA, August 1999.

Pritee Parwekar, Jaypee Institute of

21
Information Technology
 Directed Diffusion:

• In spite of SPIN, where availability of data is

advertised, in directed diffusion the BS
broadcasts interest which describes a task
required to be done by the network.
• On receiving the interest, each sensor node
then stores the interest entry in its cache and
sets up a gradient toward itself to the nodes
from which it receives the interest.

Pritee Parwekar, Jaypee Institute of

22
Information Technology
 Directed Diffusion:
• When a node has data for broadcasted
interest, it sends data through the interest’s
gradient choosing only best paths to avoid
further flooding.

Pritee Parwekar, Jaypee Institute of

23
Information Technology
 Directed Diffusion
• In a Distributed Sensor Network System, how do you
query the system for events of interest?

• Directed Diffusion is a data-centric approach.

– Two Parts
• Query Propagation
• Data Propagation

Pritee Parwekar, Jaypee Institute of

24
Information Technology
 Data-centric Protocols
• Sink sends queries to certain regions and waits
data from sensors located in that region
• Attribute-based naming is necessary to specify
properties of data

Pritee Parwekar, Jaypee Institute of

25
Information Technology
 The Directed Diffusion Protocol
• Directed Diffusion consists of several
elements:
– Interests
– Data messages
– Gradients
– Reinforcements

Pritee Parwekar,2J7aypee Institute

of
Information Technology
 Directed Diffusion - Interest & Data
• Interest
– type = wheeled vehicle
– interval = 1 s
– rect = [-100, 200, 200, 400]
– timestamp = 01:20:40 // hh:mm:ss
– expiresAt = 01:30:40
• Data
– type = wheeled vehicle // type of vehicle seen
– instance = trunk // instance of this type
– location = [125, 220] // node location
– intensity = 0.6 // signal amplitude measure
– confidence = 0.85 // confidence in the match
– timestamp = 01:20:40 // local event generation time

Pritee Parwekar,2J8aypee Institute

of
Information Technology
 Directed Diffusion - Interest Propagation

• The sink periodically

broadcasts an interest
message to each of
its neighbors.

• Every node maintains an

interest cache.

Pritee Parwekar,2J9aypee Institute

of
Information Technology
Directed Diffusion - Gradient Establishment

• That every pair of neighboring nodes establishes

a gradient toward each other.

• This technique can enable fast recovery from

failed paths or reinforcement of empirically
better paths.

Pritee Parwekar,3J0aypee Institute of

Information Technology
 Directed Diffusion - Data Propagation
• A sensor node that detects a target, it
computes the highest requested event rate
among all its outgoing gradients.

• To resend a received data message, a node

needs to examine the matching interest
entry's gradient list.

Pritee Parwekar,3J1aypee Institute

of
Information Technology
 Directed Diffusion - Reinforcement

• The node might choose

that neighbor from whom
it first received the latest
event matching the
interest to reinforce.

• It is very reactive to
changes in path quality.

Pritee Parwekar,3J2aypee Institute

of
Information Technology
Directed Diffusion:

Pritee Parwekar, Jaypee Institute of

32
Information Technology
Design Considerations

Design space for diffusion

Pritee Parwekar, Jaypee Institute of

33
Information Technology
 Directed Diffusion (Reinforcement)
• Source sends low-rate events to sink across multiple
paths
• Sink then reinforces one particular neighbor to draw
down high quality/data rate events
– Choose neighbor based on data driven rules
– Use data cache

• Sink resend the interest with higher data rate to that

reinforced neighbor.
• Results in high data rate path between source and
sink

Pritee Parwekar, Jaypee Institute of

34
Information Technology
Negative Reinforcement

Pritee Parwekar, Jaypee Institute of

35
Information Technology
 Negative Reinforcement
• The algorithm can result in more than one path being reinforced.
• If the sink reinforces neighbor A, but then receives a new event from
neighbor B, it will reinforce the path through B.
• If the path through B is consistently better (i.e., B sends events before A
does), we need a mechanism to negatively reinforce the path through A.

Pritee Parwekar, Jaypee Institute of

36
Information Technology
 Negative Reinforcement
• How to prevent multiple paths from being
reinforced??
• Idea is to negatively reinforce paths
• Choosing neighbor
– Neighbor from which no new events have come in
the last T seconds.

Pritee Parwekar, Jaypee Institute of

37
Information Technology
 Methods for Negative Reinforcement
• Soft State - time out all data gradients in the network unless they are
explicitly reinforced
e.g. With this approach, the sink would periodically reinforce neighbor B,
and
cease reinforcing neighbor A.
• Another method is to explicitly degrade the path through A by sending a
negative reinforcement message to A. This sequence of local interactions
ensures that the path through A is degraded rapidly.

Pritee Parwekar, Jaypee Institute of

38
Information Technology
 Repair
• Initiated by intermediate
nodes on a reinforced path
• Reinforce alternate path and
negatively reinforce
degraded path
• Intermediate nodes may
need to interpolate data to
prevent all downstream
nodes from initiating
reinforcement
Pritee Parwekar, Jaypee Institute of
40
Information Technology
 Loop Removal Using Negative
Reinforcement
• Negative reinforcement is also used for loop removal because
the looping paths never deliver events first.
• Truncating the looping paths for resource savings

Pritee Parwekar, Jaypee Institute of

40
Information Technology
 Loop Removal Using Negative
Reinforcement
• Loop removal is not always appropriate specifically for some shared high-
rate gradient maps with multiple sources and sinks.
• If both sources send distinguishable events, the gradient B-C and C-B
should not be truncated because each of them is necessary for delivering
events for a particular source-sink pair.
• Although such gradients may deliver some looping events, they also
consistently deliver new events. With our conservative rule for negative
reinforcement, those gradients will not be negatively reinforced.

Pritee Parwekar, Jaypee Institute of

41
Information Technology
 Rumor Routing
• Feature
– Combine query flooding and event flooding
– Discovering arbitrary paths instead of the shortest path
– Rumor routing is attractive only when
• The number of queries is larger
• The number of events is smaller

Pritee Parwekar, Jaypee Institute of

42
Information Technology
 Rumor Routing
• Basic scheme
– Each node maintain
• A lists of neighbors
• An event table

– When a node detects an event

• Generate an agent
• Let it travel on a random path
• The visited node form a gradient to the event

– When a sink needs an event

• Transmit a query
• The query meets some node which lies on the gradient
– Route establishment
Pritee Parwekar, Jaypee Institute of
43
Information Technology
 Algorithm Overview
• Assume network consists of densely distributed
wireless sensor nodes with relatively short
symmetric radio range
• Nodes records events and able to route queries
• Each node maintains a list of neighbors, events
table,
forwarding information to all the events it
knows.
• Neighbors list created and maintained by actively
broadcasting a request
 Contd…(node)
• When node witnesses an event, it adds it to its
event table, with a distance of zero to the
event.
• The basic idea of rumor routing is to use
agents to create paths leading to each event
when the event happens.
 Contd…(Agent)
• An agent is a long-lived packet
• Agent travels the network
• Propagating information about local events to
distant nodes
• Contains events table
• Synchronizes with every node it visits
• Travels network some number of hops, then dies
• Later queries can be routed along these agent-
generated paths.
Pritee Parwekar, Jaypee Institute of
48
Information Technology
 Contd…(query)
• Any nodes may generate a query and routed to a
particular event
• If node has a route to the event, it will transmit the
query
• It it does not, it will forward the query in a random
direction
• Continues until the query TTL expired, or query
reaches a node that has observed the target event
• If the node originated the query did not reach a
destination, it can always flood the query
The agent having left node B contains a route to E1
of length 3. When it arrives at node A and performs a table
sync, it will learn about the path to E2 and optimize A's
path to E1.

51
Contd…(Agents)
• After the table
synchronization
completes, the event
table will contain
the best routs to the
event
 Contd…(Agents)
• The number of agents depends on the number of event.
• Event table have expiration timestamp
• The probability of generating an agent is an algorithm
parameter
• In order to propagate directions to the event as far as
possible
in the network, a straightening algorithm is used
– The agent maintains a list of recently seen nodes.
– When picking its next hop, it will first try nodes not in the
list.
 Queries
• A query can be generated at any time by any nodes
and target to an event
• If a node has a route toward target event, it forwards
the query along the route
• If it does not, forward to random neighbor and
assume the query has not exceeded its TTL
• Query employs same mechanism as the agent, keep
list of recently seen nodes
• Some queries may not reach their destination,
application must detect the failure, flooding query
again or increase the queries TTL
 Rumor Routing
• Create paths leading to each event
• Event flooding creates a network-wide gradient field
• Query sent random walk until find the event path
• No flooding event across the network
• Query discovers event path, then route directly to
the event
• If path cannot be found, application reubmitting the
query, flooding it.
 Algorithm Stability
• Because this algorithm relies on random
decision(agent and queries)
• Performance not vary significantly over several runs
is important
 Fault Tolerance
• After the routes were established some of the
nodes were disabled
• Probability of delivery degraded slowly for 0-
20% node failure
• Over 20% node failure, the performance
degraded more severely
 Summary
• The rumor routing protocol uses a set of long-lived agents to
create paths that are directed towards the events they
encounter.
• In order to join the path, the queries are first sent on a
random walk in the network.
• Each node in the network maintains a list of its neighbors and
an event table
• If there is no route available, the node transmit a query in a
random direction.
• Then, node waits to know if the query reached the
destination for a certain amount of time, after which the node
floods the network if there is no response

Pritee Parwekar, Jaypee Institute of

62
Information Technology
 Gradient Based Routing ( GBR):
• The paradigm in GBR is calculation of a parameter,
called height of the node, which is the minimum
path between node and Base Station (BS) in terms of
the number of hops between them.
• The difference between a node’s height and the
height of its neighbor is called gradient of the link
between them.
• While forwarding data, nodes choose the links which
have largest gradient.

Pritee Parwekar, Jaypee Institute of

57
Information Technology
Gradient Based Routing ( GBR):

Pritee Parwekar, Jaypee Institute of

58
Information Technology
 Gradient Based Routing ( GBR):
• In order to prolong network life time, GBR balances the traffic
over the links using three different data distribution techniques:
• Stochastic Scheme: if there are two or more links having same
gradient, randomly choose one of them.
• Energy Based Scheme: if the available energy of the node
decreases below of a certain level, the node increases its height to
prevent other nodes sending data to it. Changing a node’s height can
also affect the heights of the node’s neighbors since a node calculates
its height to be one more than that of its lowest-height neighbor
• Stream Based Scheme: new streams are not forwarded to the
nodes which are already being a part of other streams. To achieve
this, a node that receives packets informs all its neighbors (except the
one from where the packets originate) that its height has increased.
As a consequence, the original stream remains unaffected, but new
streams may choose a different route
59
 Summary-GBR
• Memorize the number of hops when the interest
is diffused
• Minimum the number of hops to reach the BS
• To obtain balanced traffic and prolong lifetime:
– A stochastic scheme
– An energy-based
– A stream-based scheme

Pritee Parwekar, Jaypee Institute of

66
Information Technology
 Minimum Cost Forwarding Algorithm
( MCFA ):
• In MCFA each node records the estimated
least cost from itself to Base Station. Initially
at each node, the least cost is set to infinity (
∞ ).
• The BS broadcasts a message with the cost
set zero.
• Whenever a node receives the broadcast
message, compares the cost of message with
its estimated least cost.

Pritee Parwekar, Jaypee Institute of

61
Information Technology
 Minimum Cost Forwarding Algorithm (
MCFA ):
• If the estimated cost on the message plus the
cost of current link is less than what the
node has, the estimate recorded by the node
updated.
• If the estimated cost on the message plus the
cost of current link is less than what the node
has, the broadcast message is discarded.

Pritee Parwekar, Jaypee Institute of

62
Information Technology
Minimum Cost Forwarding Algorithm (
MCFA ):

Pritee Parwekar, Jaypee Institute of

70
Information Technology
 Other flat routing techniques are:
• Information-driven sensor querying and
constrained anisotropic diffusion routing
• COUGAR
• ACQUIRE
• Energy-Aware Routing

Pritee Parwekar, Jaypee Institute of

71
Information Technology
 References
• I.F. Akyildiz, W. Su, Y. Sankarasubramaniam, and E. Cayirci,
– "Wireless sensor networks: a survey".
• C. Intanagonwiwat, R. Govindan, and D. Estrin,
– "Directed Diffusion: A Scalable and Robust Communication Paradigm
for Sensor Networks“.
• David Braginsky , Deborah Estrin ,
- “Rumor Routing Algorithm For Sensor Networks “
• Zeenat Rehena , Sarbani Roy, Nandini Mukherjee ,
- “A Modified SPIN for Wireless Sensor Networks”
• Fan Ye, Alvin Chen, Songwu Lu, Lixia Zhang
- “A Scalable Solution to Minimum Cost Forwarding in Large Sensor
Networks”
• K. Akkaya, M. Younis,
– "A Survey on Routing Protocols for Wireless Sensor
Networks".
• J.N. Al-Karaki, A.E. Kamal,
– "Routing Techniques in Wireless Sensor Networks".
Pritee Parwekar, Jay7p2ee Institute of Information
Technology