Routing in nonIP Sensor Networks Mohadig Widha R' 2037179 May 2005

1
Routing in non-IP Sensor NetworksMohadig Widha
R.2037179 May 2005
2
IntroductionWireless Sensor Networks

• Progress in computing devices has lead to the
  construction of the smart sensor.
• Networks of small, cheap devices capable with
  limited memory, computation and communication
  abilities plus some simple sensors or actuators.
  In the design of such systems, energy efficiency
  and in-network processing of data are crucial
  issues.
• Properties
• Senses phenomena that happen in its environment.
• Wireless communication.
• Performs computations on data.
• Battery operated.
• Aim Draw information from the environment in
  which sensors are deployed.

3
BackgroundResource Management

• Batteries supply a limited amount of energy
• In some cases, it is difficult to recharge them
• The aim is to receive as much useful information
  as possible with a certain amount of initial
  energy provided to the sensors
• A way to reduce the energy consumption is to
  decrease the communicated information
• The production of intermediate results can reduce
  the communication cost

4
Non IP Routing Algorithms
Gossip/Rumor algorithm Ant Algorithm GEAR GPSR DCS
Directed Diffusion e3D SPIN, SLURP, PAMAS etc

• Direct
• MTE
• PEGASIS
• Static clustering
• Random clustering
• LEACH
• GRAB
• IPX

5
Basic non-ip routing

• Direct communication
• Data sent directly to the base station, naive
• Requiring a lot of transmit power
• Problem in maintenance
• MTE (Minimum Transmission Energy)
• Sent through closest intermediate node
• Only considering transmit energy neglecting
  dissipation energy at receiver node
• Necessary to go over through n transmits and n
  receives
• Required condition dAB2 dBC2 lt dAC2

6
ComparisonDirect Diffusion Communication

• Directed

• Diffusion (MTE)

7
LEACHLow-Energy Adaptive Clustering Hierarchy

• Clustering-based protocol that utilizes
  randomized rotation of cluster heads to collect
  data from neighboring nodes, aggregate data, and
  send it to the base station
• Clustering based protocol that minimizes energy
  dissipation in the networks.
• Each cluster head acts as a gateway between
  cluster members and the base station aggregating
  all the information from members, sending one
  message to the base station.
• Randomizing rotation (in specific duration of
  time) of the cluster heads to evenly distribute
  the energy among the sensor.

8
LEACH Essential Notions

• Localized coordination and control for cluster
  set-up and operation.
• Scalability
• Robustness for dynamic networks
• Local compression/aggregation to reduce global
  communication.
• Reducing energy dissipation
• Enhancing system lifetime
• Metric Power level, Location etc

9
LEACHHow does it work
10
LEACHData aggregation

• Any packet not transmitted does not need energy
• To still transmit data, packets need to combine
  their data into fewer packets ! aggregation is
  needed
• Depending on network, aggregation can be useful
  or pointless
• Relevant information of individual signal

11
LEACHAdvantage and Disadvantage

• Advantages
• Less in cost
• Realistic feasible compared with traditional
  clustering
• Global knowledge of the node position ? global
  synchronization only in certain time
• Disadvantages
• Unpredictable performance of the system
• Each node can only be cluster head temporarily
• More computation in the motes

12
e3DEnergy-efficient Distributed Dynamic Diffusion

• Utilizing location, power levels, and node load
• Distributing the energy dissipated throughout the
  network
• No data aggregation
• Very minimal overhead and realistic
• Near ideal performance 80 living in 100
  utility (synchronization issue)

13
e3D Exceptional message

• Receiving nodes queue too large
• Permanently removed from the list after some
  iterations
• Receivers power less than senders power
• Avoiding over expectance of the sender
• Receiver passing a threshold
• If the receiver almost die

14
e3D Performance
15
e3DAdvantage and disadvantage

• Advantages
• Most nodes living in the same duration
• Easy to maintenance
• No set-up phase
• When all nodes exhausted, sending process
  directly to the base station
• Disadvantage
• Higher complexity, requiring some synchronization
  messages throughout the lifetime

16
References

• Curt Schurgers, Mani B. Srivastava, Energy
  Efficient routing in WSN, NESL, UCLA, CA.
• Ioan Raicu, Efficient Even Distribution of Power
  Consumption in WSN, Department of Computer
  Science, Purdue University.
• Ioan Raicu, Scott Fowler, Loren Schwiebert,
  Sandeep K.S. Gupta. Energy-Efficient Distributed
  Dynamic Diffusion Routing Algorithm in Wireless
  Sensor Networks e3D Diffusion vs. Clustering,
  submitted for review to ACM MOBICOM 2002.
• H. Karl and A. Willig, "Protocols and
  Architectures for Wireless Sensor Networks,
  Wiley, May 2005.
• W. Heinzelman, A. Chandrakasan, and H.
  Balakrishnan, Energy-Efficient Communication
  Protocol for Wireless Microsensor Networks,
  Hawaiian Int'l Conf. on Systems Science, January
  2000.