🌐 Edge Computing Simulation

Devices on the outer ring fire requests. Each edge node handles a small local queue instantly; anything past its capacity — or a node offline — gets forwarded across the network to the central cloud, at a much higher latency.

Request rate6.0 req/s

Edge node capacity3 slots

Network congestion (cloud link)1.0×

⚠️ Simulate edge node failure: OFF

Requests/sec: 0.0
Avg. latency: 0 ms
Active edge nodes: 8 / 8

Processed at edge: 0%

Offloaded to cloud: 0%

How it works

Every request first reaches the nearest edge node (green ring = free capacity, amber/red = saturated). If the node has a free processing slot it answers locally — a short round trip. If not (or the node has failed), the request is relayed on to the central cloud and back — two extra network hops plus a heavier compute step.

edge_ok = node.active < capacity && node.enabled

t_edge  = 2·t_net_edge + t_compute_edge
t_cloud = 2·t_net_edge + 2·t_net_cloud·congestion
          + t_compute_cloud

Raise the request rate or drop the capacity and edge nodes saturate faster, pushing more amber (cloud) traffic and lifting average latency. Network congestion only stretches the cloud leg — edge trips stay cheap. Fail a node and its whole device sector routes to the cloud until it's restored.