Batch Forwarding in Wireless Sensor Networks (WSNs) Presented by Drew Wichmann Paper by Turgay Korkmaz
Wireless Sensor Networks Monitor Environment Collect and/or aggregate sensory data Send data to a base station
Forwarding Schemes Streaming Batch Forwarding Receive message, forward to next node Batch Forwarding Accumulate data, then forward to next node
Batch Forwarding Messages consist of Header (H) and Data (D) bytes Streaming Receive: m (H + D) Send: m (H + D) Batch Send: H + m D
Batch Size Static Based on known load level Dynamic Timeout-based
Advantages of Batch Forwarding Easy to concatenate data Uses contention-based MAC protocol Batch forwarding allows relay node to send most (if not all) data Lightly loaded system increases delay due to batch formation Doesn’t require aggregation
Analytical Results
Efficiency
Maximum Efficiency
Delay Performance Propagation Delay Transmission Delay Processing Delay Queuing Delay
End-to-End Delay Pattern I Streaming: Batch Forwarding: Pattern II
Average End-to-End Delay di is the end-to-end delay for data from sensor node I Pattern I: Pattern II: Average end-to-end delay Average delay for Pattern I Average delay for Pattern II
Experiments
Setup Sensor Nodes Relay Node Base Station 5 Nodes D = 12 bytes 5 Nodes Relay Node Streaming, Fixed size, and dynamic batch forwarding Fixed size: m ∈ [1,9] Dynamic: Timeout (T) = 16ms, 32ms, 64ms Buffer Size: 15, 30 messages Base Station Synchronize base station and sensor nodes Determine end-to-end delay
Fixed Batch Sizes ∆ - total size of received data Γ – total time to receive ∆ amount of data R = 250 Kbps
Fixed Batch Sizes Delivery Ratio (DR)
Fixed Batch Sizes End-to-End Delay
Dynamic Batch Forwarding Relay Node sets timer T = 16ms, 32ms, 64ms
Dynamic Batch Forwarding End-to-End Delay
Conclusion Analyzed performance of batch forwarding in WSNs Verified analysis with experimentation Best performance for dynamic batch forwarding
Questions?