Edge Data Aggregation: Sensor Bandwidth Reduction

Sensor Field

Sensor nodes 10

Sampling rate 6 Hz

Edge Node

Aggregation window 500 ms

Change-detection threshold 6%

Backhaul Link

Link capacity 8 KB/s

Live Telemetry

Raw demand

0 KB/s

Edge-forwarded

0 KB/s

Bandwidth reduction

Link utilization

Est. latency

0 ms

How it works

Each sensor samples a slowly drifting geological signal (strain, tilt, seismic amplitude) at a fixed rate and would, in a naive "send everything" design, push every reading straight to the cloud. The edge node instead buffers samples locally over an aggregation window, averages them, and forwards only when the averaged value has moved past a change-detection threshold from the last value it sent — classic delta / report-by-exception encoding used in real SCADA and IoT telemetry systems.

raw demand   = N · rate · payload           (bytes/s, every sample sent)
edge demand  = Σ forwards/window · payload  (bytes/s, only real changes sent)
reduction%   = 1 − edge demand / raw demand

utilization ρ = edge demand / link capacity
queue delay   = propagation + ρ/(1−ρ) · service_time   (M/M/1 waiting time)

Sensor nodes / Sampling rate — how much raw data the field generates; more sensors or a higher rate raises both raw and edge demand.
Aggregation window — how long the edge node holds samples before deciding whether to report; longer windows smooth noise but delay detection.
Change-detection threshold — how large a swing is needed before the edge bothers forwarding; raise it to cut bandwidth further at the cost of missing small changes.
Link capacity — the bandwidth-limited backhaul from the remote site to the cloud. When edge-forwarded demand exceeds it, packets queue and the link is marked saturated.

This is the actual resource-allocation trade-off edge computing solves: local processing near the data source cuts the volume that has to cross a slow, expensive, or unreliable long-haul link, at the cost of some detection latency and local compute.

LINK OK

Sensor → edge (raw sample) Edge → cloud (forwarded)

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.