Edge-Fog-Cloud Latency Simulator

Simulate request-response latency across edge, fog, and cloud tiers.

Learner-ready edge-fog-cloud latency simulator with animated packet paths, deterministic latency formulas, scenario presets, tier comparisons, bottleneck diagnosis, and mobile-safe learning support.

Edge Fog Cloud Latency Budget

Edge-Fog-Cloud Latency Simulator

Send the same IoT request to three placement tiers and watch where the time goes. The model separates data transfer, physical distance, queueing, and compute so learners can see why the closest tier is not always the best tier, but distance dominates tight control loops.

Edge Fastest tier now

1 of 3 meet Budget result

Cloud 18x edge Latency spread

Propagation Slowest-tier bottleneck

Distance is paid twice

A request-response loop includes the trip to the tier and the command or result back to the device.

Bandwidth only helps transfer

Faster links shrink payload time but do not remove WAN propagation delay.

Compute can flip a result

Heavy analytics may favor fog or cloud when edge hardware cannot process fast enough.

Budget decides fit

A tier can be slower and still acceptable when the application budget is loose.

Scenario Presets

Each preset changes payload, workload, network, compute, and response budget together.

Playback

Play compares all three tiers on one normalized timeline. Edge can finish while fog and cloud are still moving.

Workload

Payload size 64 KB

Processing work Vision crop

Response budget 20 ms

Queue load Light

Network Profile

Plant LAN: local fog is near the device, while cloud still crosses a WAN path.

Compute Capacity

Edge compute AI module

Fog compute Gateway GPU

Cloud compute GPU node

Focus Tier

Request-Response Timeline

The same event is sent to edge, fog, and cloud. The token returns when that tier has completed the loop.

0.0 ms elapsed on comparison timeline

IoT Device 64 KB event, 1.3 KB command

Edge Runtime Local transfer and local compute

Fog Node 2 hops through plant LAN

Cloud Region WAN round trip plus cloud compute

Edge 0.0 ms

Ready at device.

Fog 0.0 ms

Ready at device.

Cloud 0.0 ms

Ready at device.

Latency Results

Edge Processing 10.5 ms Meets 20 ms

Fog Propagation 24.8 ms Close to 20 ms

Cloud Propagation 188 ms Misses 20 ms

Edge

10.5 ms

Fog

24.8 ms

Cloud

188 ms

Capture Transfer Propagation Queueing Processing

Recommended: Edge

Edge is the only tier that meets the current 20 ms control-loop budget. Cloud compute is faster, but the WAN round trip dominates the total.

• Use edge for immediate control decisions.
• Send summaries to fog or cloud after the local action is complete.

Edge formula

The focused tier uses the same deterministic request-response model as the others.

Total = capture + transfer + propagation + queueing + processing

How The Model Works

Total latency is a round-trip teaching estimate:

• Capture: fixed device event preparation time.
• Transfer: request payload plus return command serialized over the selected link.
• Propagation: distance and hop delay for the request and response path.
• Queueing: deterministic load penalty based on path length and selected queue load.
• Processing: workload units divided by the selected compute capacity.

Practice Checks

1. Set the budget to 10 ms. Which tier still works, and why?
2. Increase payload size. Which component grows first?
3. Switch from Plant LAN to Congested. Which tier is hurt the most?
4. Use Fleet analytics. Why can a high-latency tier still be acceptable?

Technical Notes

The transfer calculation uses 1 KB = 1024 bytes and Mbps = 1,000,000 bits per second. The simulator is deterministic on purpose: queueing is a selected load penalty, not random jitter, so learners can compare settings reliably.