1. On the Lifetime of Wireless Sensor Networks
Isabel Dietrich
Falko Dressler
ACM Transaction on Sensor Networks, 2009
Presented by Group 6
Tianyi Pan

2. Why Limited Lifetime for WSNs?
Various application domains
Wildlife monitoring
Precision agriculture
Logistics applications
Environmental conditions

3. Outline of the Paper Analysis of existing lifetime definitions
Overview of the parameters influencing network lifetime
Concise redefinition of network lifetime
Application of the definition

4. Existing Lifetime Definitions
Based on the number of alive nodes.
Based on sensor coverage.
Based on connectivity.
Based on application quality of services requirements
Based on combined measures

5. Lifetime Based on the Number of Alive Nodes
n-of-n lifetime 𝑇 𝑛 𝑛 = min 𝑣∈𝑉 𝑇 𝑣
Normally too short and neglects redundancy
Not capable with hardware failures
k-of-n lifetime 𝑇 𝑛 𝑘 : at least k out of n nodes are alive
Cannot cope with critical nodes
m-in-k-of-n: at least k out of n nodes alive and none of the m critical nodes fail
The time till the first cluster head fail
Require static clusters, which is unrealistic
Time till all nodes fail
Too optimistic

6. Lifetime Based on Sensor Coverage
Time till certain coverage requirement cannot be met
Area/volume coverage
Target coverage
Barrier coverage
Detect all trespassing activities in a region
𝛼-coverage
Only 𝛼 percent of the region of interest must be covered (by at least 1 sensor)
k-coverage
Each point of interest in the region must be covered by at least k sensors.

7. Lifetime Based on Connectivity
The minimum time that the size of the largest connected component drops to a certain size.
The minimum time that the percentage of nodes having paths to the sink drops to a certain value.
Total number of packets transmitted to the sink.
Dependent on algorithms (variation in # packets)
Cannot clearly indicate the lifetime
Number of successful data gathering trips

8. Lifetime Based on Application Quality of Services Requirements
The time period during which the WSN continuously satisfies the application requirement
Too abstract, application characteristics unspecified
The time until the network no longer provides an acceptable event detection ratio
A specific application requirement yet still vague

9. Lifetime Based on Combined Measures
Coverage & Connectivity
The time when either drops below a threshold
Cardinality of the largest connected component, number of alive nodes, 𝛼-coverage
The time when any of the three is unsatisfied
Share the issues with single aspect
Can omit some aspects

10. Overview of the Parameters
Node mobility and topology changes
Heterogeneity
Application characteristics
Quality of service

11. Node Mobility and Topology Changes
Mobility happens in all WSNs
Node failure in static WSN (similar to mobility)
Sensors moved by external forces
Mobility can complicate analysis of lifetime
Connection can be lost and then recover
System “die” and resurrect
Mobility can improve sensor coverage
Act as relay
Mobile sinks

12. Heterogeneity Varying battery power and power consumption
Cluster head may have much larger energy reservoir
Some nodes may need to send more data
Some location may need to be covered by different types of sensor
Transmission ranges can vary
Different mobility patterns

13. Application Characteristics
Sensors carry different tasks of an application
Need to work in synergy
Destination for data packets can affect communication patterns (with multiple sinks)
Node activity level
Activity can be triggered by:
Events
Regular time intervals
Other nodes

14. Quality of Service Coverage Event detection ratio Exposure
Connectivity (availability, latency, loss)
Requirements on continuous service
Observation accuracy
Main QoS parameters for WSNs

15. The Concise Definition of Lifetime
Basic notations
Set of sensor types Y
Set of all sensor nodes 𝑆 𝑌 , 𝑆 𝑌 =𝑛
Set of alive nodes
Set of active nodes
Set of active nodes (relaxed)
Set of sink nodes 𝐵 𝑡 ⊂ 𝑆 𝑌

16. Basic Notations Communication graph 𝐺 𝑡 =(𝑉 𝑡 ,𝐸(𝑡))
Ability of two nodes to communicate in the time interval [𝑡−Δ𝑡,𝑡]
Set of target points
Sensing area for type 𝑦

17. Relax the Hard Constraints
Graceful degradation
Quality drops when hit a soft limit, before hitting the hard limit
No “sudden death”
𝜁≥1 means perfect fulfilment
𝜁<0 means no fulfilment
In between indicates the service level
Time-integrated criteria
Requirement can be fulfilled (at least once) within time intervals, not at every time point.

18. The Criteria
Alive nodes: the portion of alive nodes in all existing nodes 𝜓 𝑙𝑛 𝑡 = 𝑢 𝑡 𝑛
Latency: the portion of packets having shorter delay that pre-specified max latency 𝑙
Delivery ratio: portion of packets received correctly

19. The Criteria Connectivity
Indicator for node 𝑣 having connection to any active sink node in 𝐵(𝑡)
The portion of active nodes having a connection to a sink

20. The Criteria
Area coverage: fraction of region covered by type-y sensors
Target coverage: fraction of targets covered by type-y sensors
𝒌-coverage: single read may be inaccurate

21. The Criteria Barrier coverage
Detect and track all intruder in the region
Sensor must cover all possible paths the intruder may take

22. The Criteria Connected coverage
Required in need of transmitting the data of any point of interest to a sink at any time
Coverage only possible via the sensor nodes that are connected to a sink

23. Network Lifetime System up/down times Service disruption
Each working interval
Accumulated network lifetime 𝒁 𝒂 = 𝒊=𝟎 𝒆 𝒕 𝒊 𝒂
Total network lifetime 𝒁 𝒕 = 𝒕 𝒆
Allowed disruption

24. Mapping of Existing Definitions

25. Mapping of Applications

26. Mapping of Applications
Habitat monitoring
Sensors need to have connectivity to the sink at least once per time interval (a day)
Allow outages for some time (a day)
Not critical, not coverage

27. Mapping of Applications
Intrusion detection
Connected barrier coverage
Short delay and limited packet loss
No service disruption
Critical and coverage

28. Mapping of Applications
Smart buildings
Intermittent coverage of the whole area
Send data to sink once every time interval
Short disruption allowed
Coverage, not critical

29. Mapping of Applications
Human physiological data
Need continuous connection to the sink
Short delay and limited packet loss’
No service disruption
Critical but not coverage

30. Questions?