What does this simulation show?

It removes nodes from a network one at a time and tracks how connectivity collapses, measuring the giant component — the largest set of still-connected nodes — after every removal.

Why do scale-free and random networks react differently to attack?

A scale-free (Barabási–Albert) network depends on a few high-degree hubs, so removing them by degree fragments it fast. An Erdős–Rényi random network has no dominant hubs, so targeted and random removal cause similar, gradual damage.

Home▸Networks & Graph Theory▸Network Resilience (2D)

Network Resilience (2D)

A flat-canvas companion to the 3D Network Resilience lab: build a Barabási–Albert scale-free graph or an Erdős–Rényi random graph, then attack it node by node — targeting the highest-degree hub each time, or picking victims at random — and watch the giant component either survive or collapse.

Networks & Graph Theory2DModerate60 FPS📱 Mobile-adapted⇄ 3D version

Details

CategoryNetworks & Graph Theory

RenderingCanvas 2D

DifficultyModerate

Target FPS60

ControlsNodes slider (20-150) sets network size. Network-type tabs switch between Scale-Free (Barabási–Albert) and Erdős–Rényi random graphs. Attack-mode tabs switch between Targeted (highest-degree hub first) and Random removal. A removal-speed slider (1×-5×) sets nodes removed per frame. Attack/Pause toggles automatic removal, Step removes one node at a time, and Rebuild regenerates the network.

barabási–albertscale-free networkshub attackgiant componentpercolationerdős–rényi

Related simulations

2D Force-Directed GraphBuild and edit a network live on a flat 2D canvas driven by real force physics.

Epidemic Spread on a Random Network (2D)A companion network-science simulation on a flat canvas.

Network Resilience (3D)The WebGL original this page is a Canvas 2D companion to. ⊞ All 154 in Networks & Graph Theory →

This 2D companion runs the same graph-percolation mechanic as the 3D version on a plain canvas: a Barabási–Albert scale-free network is grown by preferential attachment, or an Erdős–Rényi random graph is generated with uniform edge probability, and a force-directed layout keeps the drawing readable as nodes are removed. Each removal — hub-first in targeted mode, uniformly random otherwise — triggers a breadth-first search that recomputes connected components, so the giant-component size and a live mini history chart update after every step, making the "robust yet fragile" asymmetry of scale-free networks directly visible.

⚙ Under the hood

2D canvas percolation lab: Barabási–Albert scale-free and Erdős–Rényi random graphs, targeted-hub vs random node removal, and a BFS-recomputed giant component after every step.

barabási–albertscale-free networkshub attackgiant componentpercolationerdős–rényi

2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install