SYN Flood Backlog Queue (2D)
An Erlang-B loss-system model of a TCP SYN backlog: watch a live 2D slot grid and occupancy chart, and compare the measured backlog against an independently computed analytic prediction, with and without SYN cookies.
This is a queueing-theory view of the same TCP SYN-flood vs SYN-cookies phenomenon: instead of an orbiting 3D scene, the backlog is drawn as a flat grid of slots with a live occupancy strip-chart underneath. A real discrete-event simulation fires legitimate and spoofed attack SYNs at the rates you choose, fills slots on arrival, and expires them after the half-open timeout (or instantly on a successful legit handshake) — while, completely independently, a birth-death Markov-chain solver computes the exact steady-state occupancy and blocking probability that classical queueing theory predicts for the same arrival and service rates. Watch the two agree: the measured backlog curve settles onto the dashed analytic prediction, and flipping on SYN cookies collapses both to zero regardless of attack volume.
2D loss-system model of a TCP SYN backlog: a live slot grid and occupancy strip-chart driven by a real discrete-event packet simulation, cross-checked frame by frame against an independently computed Erlang-B analytic prediction (mean occupancy and blocking probability), with a SYN cookies toggle that collapses the backlog to zero.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install