TryIncrease Packets per second until utilisation passes 100%.
ObserveWatch Goodput, p95 latency, and Queue delay estimate after pressing Calculate.
ExplainExplain why latency rises when utilisation passes 100%.
Technical boundaries. Queue delay is estimated from utilisation. Goodput uses packet loss and link capacity, while p95 latency uses average latency and jitter range.
Test Before Hardware
Start here before building the ESP32 version. The controls below let you model the same performance ideas directly in the browser: bandwidth, offered load, latency, jitter, packet loss, and goodput. Begin with offered traffic comfortably below link capacity and record utilization, average latency, jitter range, loss, and delivered goodput. Then raise packet rate or packet size until utilization crosses the congestion region, observing which metric changes first. Finally vary base latency and loss independently so propagation, queueing, and delivery failure are not confused. Carry the same predicted thresholds into the hardware run and explain any difference with measured packet timing or implementation overhead.
NoteTry It: Network Performance Workbench
viewof perfLabBandwidth = Inputs.range([10,5000], {value:1000,step:10,label:"Link bandwidth (kbps)"})viewof perfLabPacketSize = Inputs.range([20,1000], {value:120,step:10,label:"Packet size including headers (bytes)"})viewof perfLabPacketRate = Inputs.range([1,2000], {value:500,step:1,label:"Packets per second"})viewof perfLabBaseLatency = Inputs.range([5,300], {value:35,step:5,label:"Base latency (ms)"})viewof perfLabJitter = Inputs.range([0,120], {value:20,step:1,label:"Jitter range (+/- ms)"})viewof perfLabLoss = Inputs.range([0,30], {value:2,step:1,label:"Packet loss (%)"})