If you process large files or gigabyte-scale logs in Node.js, you've likely hit the same wall: scanning through a massive file either crashes the runtime with JavaScript heap out of memory or completely freezes the single-threaded event loop.
To solve this at the systems level, I built FastScan Ultra — a hardware-accelerated native C addon designed for raw throughput and zero-latency execution.
Key architecture highlights:
• Zero-Copy Virtual Memory: Maps files directly using kernel primitives (mmap / MapViewOfFile). Data never enters the V8 heap (0 MB heap allocation).
• SIMD Vectorization: Saturates modern CPU pipelines (AVX-512, AVX2, and NEON) with dual-byte prefiltering, outperforming standard Node.js streaming by 2x to 3.4x in real storage benchmarks.
• Non-Blocking Thread Pool: Offloads scans to dedicated native worker threads with sub-microsecond wakeups, keeping HTTP event loops 100% responsive.
• Multi-Pattern Forensics: Matches hundreds of threat indicators (IoCs) or error signatures in a single continuous pass through memory.
Fully open-source under MIT:
🔗 GitHub Repository: https://github.com/okba14/FastScan
Feedback, pull requests, and benchmarks on your own infrastructure are welcome. For technical discussions or collaborations, reach out directly at:
📧
contact@guiarx.com

Top comments (0)