PLEX Runtime™ – Pulsed Layered Exhaustive eXecution

A deterministic runtime for exhaustive compute workloads.
Non-commercial MIT; commercial license required for monetized use.

Table of Contents

1. Overview
2. Why PLEX Exists
3. Key Concepts
4. Execution Model
5. Installation
6. CLI Usage
7. Reference Implementation
8. Dual Licensing
9. Contributing
10. Contact

Overview

PLEX Runtime™ is a reference execution engine for exhaustive workloads that require:

• Every byte examined
• Deterministic execution
• Linear, verifiable scaling
• No sampling, no shortcuts

Target Workloads:

• Security and compliance scans
• Full-dataset pattern matching
• Verification and audit workloads
• Broad anomaly detection

PLEX is not a database, query language, or framework — it is a formal execution model with correctness-first scheduling.

Why PLEX Exists

Modern data lakes and lakehouses store petabytes of unstructured and semi-structured data. Traditional indexed engines fail for workloads that touch:

• 8–10% of total data

• Large portions without selective keys

Problems include:

• Indexes collapse
• Latency becomes unpredictable
• Costs explode
• Coverage is incomplete

PLEX is designed for deterministic exhaustive scanning, enabling predictable costs, guaranteed coverage, and linear scaling.

Key Concepts

PLEX Runtime™ enforces four core invariants:

1. No shared mutable state
2. No global synchronization
3. Pulsed work dispatch
4. Deterministic per-unit execution

Benefits:

• Linear scaling from single-node to multi-cluster
• Predictable saturation
• Deterministic, reproducible results

Execution Model

PLEX decomposes work into a layered fan-out execution tree:

Controller
├─ Partition Layer
│   ├─ Chunk Layer
│   │   ├─ Scan Units

Pulse Behavior:

• Work is released in bounded pulses
• Each pulse is independently verifiable
• Pulses prevent system overload

This contrasts with traditional map-reduce or unbounded task queues.

Installation

Node.js / NPM

npm install @iconoclast/plex-runtime

From Source

git clone https://github.com/jamesbrianchapman/PLEX-Runtime.git
cd PLEX-Runtime
npm install

CLI Usage

After installation, the plex CLI is available:

# Run exhaustive scan on dataset
plex run ./data.json --pulse-size 32 --max-concurrency 8

Options:

• --pulse-size: Number of tasks per pulse
• --max-concurrency: Maximum concurrent tasks per node
• --help: Display CLI help

Reference Implementation

The repository includes a non-commercial reference runtime demonstrating:

• Pulsed scheduling
• Deterministic task execution
• Task isolation guarantees

Example usage:

const PlexRuntime = require('@iconoclast/plex-runtime');

const runtime = new PlexRuntime({ pulseSize: 32, maxConcurrency: 8 });

(async () => {
  const dataset = [...]; // Array of data items
  const results = await runtime.run(dataset, async (item) => {
    // Example scan function
    return item.value * 2;
  });
  console.log(results);
})();

Note: This implementation does not include GPU optimizations, peak saturation tuning, or proprietary pulse shaping algorithms — those are reserved for commercial licensees.

Dual Licensing

PLEX Runtime™ uses a dual license model:

1. MIT License – Non-Commercial Use

• Research, experimentation, local development
• Academic or personal projects
• Prohibited for revenue-generating use

2. Commercial License – Required for Monetized Use

• Any paid service or internal revenue-generating use
• Includes optimized runtime, GPU backends, and enterprise support
• Contact for licensing: [email protected]

Contributing

Contributions are welcome for:

• Bug fixes
• Documentation
• Non-commercial runtime improvements

All contributions require a Contributor License Agreement (CLA). See CLA.md for details.

Contact

Owner / Maintainer: James Brian Chapman Emails for verification / communication:

• [email protected]
• [email protected]
• [email protected]
• [email protected]

GitHub: https://github.com/jamesbrianchapman/PLEX-Runtime

Disclaimer: This repository discloses only the minimal execution logic for research and non-commercial use. Proprietary peak-performance optimizations remain reserved for commercial licensees.