Physics-based
concurrency control
Replaces static rate limits with adaptive thresholds, deterministic backpressure, and priority-based load shedding.
586M
ops/second (WASM)
2.11ns
latency per decision
184
tests passing
The Core Equation
R(t) = P + M + S
The Physics Engine
Every route has resistance that changes based on real-time telemetry. Traffic routing as flow through impedance networks.
Adaptive Thresholds
Eliminates manual configuration through Z-Score based auto-tuning. The system learns your traffic patterns.
dynamicBreak = μ(R) + 3σ(R) Momentum Detection
Derivative-based resistance detects problems before they peak. Reduces state transitions by 98%.
M = dP/dt Scar Tissue Memory
Historical trauma remembers bad routes. Automatic load redistribution to healthy nodes.
S(t) = S(t-1) × e^(-λt) + trauma Quick Start
Get up and running in under a minute. No configuration required.
npm install atrionWhat you get:
- Rust/WASM physics engine (586M ops/sec)
- TypeScript fallback for compatibility
- Zero-config auto-tuning out of the box
- Full TypeScript types included
import { Atrion } from 'atrion'
// Initialize with auto-tuning
const atrion = new Atrion()
await atrion.connect()
// Make routing decision
const decision = atrion.route('api/checkout', {
latencyMs: 45,
errorRate: 0.01,
saturation: 0.3,
})
if (!decision.allow) {
return res.status(503).json({
error: decision.reason
})
}
// decision.resistance = current Ω
// decision.mode = 'OPERATIONAL' | 'CIRCUIT_BREAKER' Pro tip: The engine automatically selects
WASM when available. Use engine: 'auto' for optimal performance.