1. A Survey on Routing Protocols for Wireless Sensor Networks
By Kemal Akkaya & Mohamed Younis
Presented by
Aggelos Vlavianos
Jorge Mena
Department of Computer Science and Engineering
02/08/05

2. Table of Contents Goals Introduction System Architecture & Designing
Protocols for Sensor Networks
Summary

3. Goals A survey of recent routing protocols for sensor networks
Classification of these protocols

4. Introduction Sensors are micro-electro-mechanical systems (MEMS)
Low power devices
Data processing capable
Communication capabilities

5. Introduction - Usage
Gather data locally (Temperature, Humidity, Motion Detection, etc.)
Send them to a command center (sink)
Applications
Surveillance
Security
Disaster Management
Environmental Studies

6. Introduction - Constraints
Limitations
Energy Constrains
Bandwidth
All layers must be energy aware
Need for energy efficient and reliable network routing
Maximize the lifetime of the network

7. Introduction - Routing
No global addressing
Redundant data traffic
Multiple-source single-destination network
Careful resource management
Transmission power
On-board energy
Processing capacity
Storage

8. System Architecture & Designing
Network Dynamics
Mobile or Stationary nodes
Static Events (Temperature)
Dynamic Events ( Target Detection)
Node Deployment
Deterministic – Placed manually
Self-organizing – Scattered randomly

9. System Architecture & Designing
Energy Considerations
Direct vs Multi-hop communication
Direct Preferred – Sensors close to sink
Multi-hop – unavoidable in randomly scattered networks
Data Delivery Models
Continuous
Event-driven
Query-driven
Hybrid

10. System Architecture & Designing
Node Capabilities
Homogenous
Heterogeneous
Nodes dedicated to a particular task (relaying, sensing, aggregation)
Data Aggregation/Fusion
Aggregation – Combination of data by eliminating redundancy
Data Fusion is Aggregation through signal processing techniques
Aggregation achieves energy savings

11. Introduction - Taxonomy
Classification of Routing Protocols
Data Centric
Hierarchical
Location-based
Network Flow & QoS Aware

12. 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

13. Data-centric Protocols
Flooding
Gossiping
Sensor Protocols for Information via Negotiation (SPIN)
Directed Diffusion
Energy-aware Routing
Rumor Routing
Gradient-Based Routing (GBR)
Constrained Anisotropic Diffusion Routing (CADR)
COUGAR
ACtive QUery forwarding In sensoR nEtworks (ACQUIRE)

14. Data-centric Protocols
Flooding
Sensor broadcasts every packet it receives
Relay of packet till the destination or maximum number of hops
No topology maintenance or routing
Gossiping
Enhanced version of flooding
Sends received packet to a randomly selected neighbor

15. Data-centric Protocols – Flooding, Gossiping Problems
Problems of Implosion, Overlap, Resource Blindness

16. Data-centric Protocols
Sensor Protocols for Information via Negotiation (SPIN)
Uses meta data and high level descriptors to name the data

17. Data-centric Protocols - SPIN
Topological changes are localized - Each node needs to know only its neighbors
SPIN halves the redundant data in comparison to flooding
Cannot guarantee data delivery
SPIN NOT good for applications that need reliable data delivery

18. Data-centric Protocols
Directed Diffusion
Uses a naming scheme for the data to save energy
Attribute-value pairs for data and queries on-demand (Interests)
Interests are broadcasted by the sink (query) to its neighbors (caching), which can do in-network aggregation
Gradients = reply links to an interest (path establishment)

19. Data-centric Protocols – Direct Diffusion
Energy saving and delay done with caching
No need for global addressing (neighbor-to-neighbor mechanism)
Cannot be used for continuous data delivery or event-driven applications

20. Data-centric Protocols
Energy-aware Routing
Occasional use of a set of sub-optimal paths
Multiple paths used with certain probability
Increase of the total lifetime of the network
Hinders the ability for recovering from node failure
Requires address mechanism Complicate setup
The optimal path is not used all the time ….

21. Data-centric Protocols
Rumor Routing
Variation of Directed Diffusion
Flood the events instead of the queries
Creation of an event  generation of a long live packet travel through the network (agent)
Nodes save the event in a local table
When a node receives query  checks its table and returns source – destination path

22. Data-centric Protocols – Rumor Routing
Advantages
Can handle node failure
Significant energy savings
Disadvantages
Works well only with small number of events
Overhead through adjusting parameters, like the time to live of the agent
Big number of events mean event tables to big and big cost for maintaining agents

23. Data-centric Protocols
Gradient-Based Routing (GBR)
Slightly changed version of Directed Diffusion
Keep the number of hops to the sink when an interest is created (height of the node)
Node’s height – neighbors height = gradient of the link
Node forward packet to the link with largest gradient

24. Data-centric Protocols -GBR
Traffic Spreading and Data Aggregation  Balance uniformly the network traffic
Stochastic scheme
Energy-based scheme
Stream-based scheme
Outperforms Directed Diffusion in terms of total communication energy

25. Data-centric Protocols
Constrained Anisotropic Diffusion Routing (CADR)
General form of Directed Diffusion
Query Sensors
Route data in the network
Activates sensors close to the event and dynamically adjusts routes
Routing based on a local information/cost gradient
More energy efficient than Directed Diffusion

26. Data-centric Protocols
COUGAR
Views the network as a huge distributed database
Declarative queries to abstract query processing from network layer functions
Introduces a new query layer
Leader node performs data aggregation and transmits to the sink

27. Data-centric Protocols - COUGAR
Disadvantages
Additional query layer brings overhead in terms of energy consumption and storage
In network data computation requires synchronization (i.e. wait for all data before sending data)
Dynamically maintenance of leader nodes to prevent failure

28. Data-centric Protocols
ACtive QUery forwarding In sensoR nEtworks (ACQUIRE)
Views network as a distributed database
Node receiving a query from the sink tries to respond partially and then forwards packet to a neighbor
Use of pre-cached information
After the query is answered, result is returned to the sink by using the reverse path or the shortest path
If cache information is not up to date  node gathers information from neighbors within look ahead of d hops

29. Data-centric Protocols – ACQUIRE
Motivation: Deal with one shot complex queries
Efficient routing by adjusting parameter d
If d equals network size  behaves similar to flooding
If d too small the query has to travel more hops

30. Hierarchical Protocols
Maintain energy consumption of sensor nodes
By multi-hop communication within a particular cluster
By data aggregation and fusion  decrease the number of the total transmitted packets

31. Hierarchical Protocols
LEACH – Low-Energy Adaptive Clustering Hierarchy
Power-Efficient GAthering in Sensor Information Systems (PEGASIS)
Hierarchical PEGASIS
Threshold sensitive Energy Efficient sensor Network protocol (TEEN)
Adaptive Threshold TEEN (APTEEN)
Energy-aware routing for cluster-based sensor networks
Self-organizing protocol

32. Hierarchical Protocols
LEACH – Low-Energy Adaptive Clustering Hierarchy
One of the first hierarchical routing protocols
Forms clusters of the sensor nodes based on received signal strength
Local cluster heads route the information of the cluster to the sink
Cluster heads change randomly over time  balance energy dissipation
Data processing & aggregation done by cluster head

33. Hierarchical Protocols - LEACH
Advantages
Completely distributed
No global knowledge of the network
Increases the lifetime of the network
Disadvantages
Uses single-hop routing within cluster not applicable to networks in large regions
Dynamic clustering brings extra overhead (advertisements, etc)

34. Hierarchical Protocols
Power-Efficient GAthering in Sensor Information Systems (PEGASIS)
Improvement of LEACH
Forms chains from sensors rather than clusters
Data aggregation in the chain  one node sends the data to the base station
Outperforms LEACH
Excessive delay for distant nodes in the chain
Outperforms LEACH  we don’t have the overhead of the cluster formation

35. Hierarchical Protocols
Hierarchical PEGASIS
Extension of PEGASIS
Decrease the delay for the packets during transmission to the base station
Solution to the delay data gathering problem
Simultaneous transmissions of data messages
Avoid collisions and possible signal interference
Signal Coding (e.g. CDMA)
Spatially separated nodes can transmit at the same time

36. Hierarchical Protocols
Threshold sensitive Energy Efficient sensor Network protocol (TEEN)
Good for time-critical applications
Hierarchical along with a data-centric approach
Hierarchical grouping: Close nodes form clusters and this process goes on the second level until sink is reached
Cluster headers broadcast:
Hard Threshold
Soft Threshold
Not good for applications that need periodic reports

37. Hierarchical Protocols

38. Hierarchical Protocols
Adaptive Threshold TEEN (APTEEN)
Captures both periodic data collection and reacting to time-critical events
APTEEN supports queries:
Historical -Analyze past data values
One-Time –Take a snapshot of the current network view
Persistent monitor an event for a period of time

39. Hierarchical Protocols – TEEN & APTEEN
Advantages
Outperform LEACH in terms of energy dissipation and total lifetime of the network
Disadvantages
Overhead and complexity of:
Forming multiple level clusters
Implementing threshold-based functions
Dealing with attribute-based naming of queries

40. Hierarchical Protocols
Energy-aware routing for cluster-based sensor networks
Assumptions:
Sensors are grouped into clusters prior to network operation
Cluster Heads (Gateways) less energy constrained
Cluster Heads know the location of the sensors  Known Multi-Hop routing to collect data
Communication node (sink) communicates only with gateways

41. Hierarchical Protocols
Stages of a Sensor inside a cluster
Sensing only
Relaying only
Sensing-Relaying
Inactive

42. Hierarchical Protocols
Least Cost path used between nodes and gateway
Cost function
Energy Consumption
Delay Optimization
Performance Metrics
TDMA based MAC is used for nodes to send data to the gateways
Protocol performs well for
Energy-based metrics (e.g. network lifetime)
Cotemporary metrics (e.g. throughput)

43. Hierarchical Protocols
Self-organizing protocol
Architecture supports heterogeneous sensors
Nodes act as routers  backbone of communication
Stationary
Sensing nodes forward data to the routers
Mobile
Sensors are a part of the network if they are reachable by a router

44. Hierarchical Protocols
The architecture requires addressing
Sensor identified by the router is connected to
Algorithm for Self-Organizing & Creation of routing table
Discovery phase
Organization phase
Maintenance phase
Self-reorganization phase
Utilizes router nodes to keep all sensors connected by forming a dominating set

45. Hierarchical Protocols
Advantages
Useful for applications which need communication of a specific node (e.g. parking-lot networks)
Small cost of maintaining routing tables
Keeping routing hierarchy strictly balanced
Energy Savings – Utilization of a limited subset of nodes
Disadvantages
Organization phase not on demand
Many cuts in the network increase the probability of applying reorganization phase

46. Location-based Protocols
Distance between two nodes is calculated using location information
Energy consumption can be estimated
Efficient energy utilization
Protocols designed for Ad hoc networks with mobility in mind
Applicable to Sensor Networks as well
Only energy-aware protocols are considered

47. Location-based Protocols
MECN & SMECN
Minimum Energy Communication Network
GAF
Geographic Adaptive Fidelity
GEAR
Geographic and Energy Aware Routing

48. MECN & SMECN Utilizes low power GPS
Best applicable to non-mobile sensor networks
Identifies a relay region
Find a sub-network
to relay traffic
Self-reconfiguring
Dynamically adaptive
Minimum Energy Communication Network
Small Minimum Energy Communication Network

49. SMECN Assumption: broadcasting allowed Obstacles considered
Smaller network (number of edges)
Less hops per transmission
More energy efficient than MECN
Less maintenance cost of links
More overhead introduced

50. GAF Forms a virtual grid for covered area
Nodes use GPS to associate itself to the grid
Nodes are allowed
to be turned off if are
equivalent.
Handle mobility

51. GAF Three States Representative Node is always active
Discovery
Active
Sleep
Representative Node is always active
No aggregation or fusion performed
As good as a normal Ad hoc in terms of latency and packet loss (saving energy)

52. GEAR Uses heuristics to route packets
Energy aware neighbor
Geographic informed neighbor
Same as Direct Diffusion but restricts interest by region
Estimated Cost
Residual energy and distance to destination
Learned Cost
Accounts for routing around holes
Geographic and Energy Aware Routing
Hole: occurs when a node does not have any closer neighbor to the target region than itself

53. Network Flow & QoS-aware Protocols
Network Flow: Maximize traffic flow between two nodes, respecting the capacities of the links
QoS-aware protocols consider end-to-end delay requirements while setting up paths

54. Network Flow & QoS-aware Protocols
Maximum Lifetime Energy Routing
Maximum Lifetime Data Gathering
Minimum Cost Forwarding
Sequential Assignment Routing
Energy Aware QoS Routing Protocol
SPEED

55. Maximum Lifetime Energy Routing
Maximizes network lifetime by defining link cost as a function of:
Remaining energy
Required transmission energy
Tries to find traffic distribution (Network flow problem)
The least cost path is one with the highest residual energy among paths

56. Maximum Lifetime Data Gathering
Maximizes the Data-gathering schedule
Maximum Lifetime Data Aggregation
Data aggregation plus setting up maximum lifetime of routes
Maximum Lifetime Data Routing
When data aggregation is not possible
Computational Expensive (scalability)
Clustering MLDA

57. Minimum Cost Forwarding
Aims at finding the minimum cost path in a large network, simple and scalable
Cost function captures delay, throughput, and energy metrics from node to sink
Back-off based algorithm
Finds optimal cost of all nodes to the sink by using only one message per node
Does not require addressing or forwarding paths

58. Sequential Assignment Routing
Table-driven, multi-path protocol
Creates trees rooted at immediate neighbors of the sink (multiple paths)
QoS metrics, energy resource, priority level of each packet
Failure recoverable (done locally)
High overhead to maintain tables and states at each sensor

59. Energy Aware QoS Routing Protocol
Finds least cost and energy efficient paths that meet the end-to-end delay during connection
Energy reserve, transmission energy, error rate
Class-based queuing model used to support best-effort and real-time traffic

60. Energy Aware QoS Routing Protocol

61. SPEED
Each node maintains info about its neighbors and uses geographic forwarding to find the paths
Tries to ensure a certain speed for each packet in the network
Congestion avoidance

62. Summary

63. Questions