TECHNICAL PAPERS

Papers

In-depth technical papers on packet intelligence architecture. Each paper is the full version — the blog articles are the distilled form.

Download PDF
Engineering Deep Dive 35 min read
Why Querying 100 EB Takes Seconds
The architecture behind Vantage Query — and why it has nothing to do with a database
0.000019% Data Read Per Query 8× Window Scaling 0.021% Index Overhead PCAPNG Compatible
A targeted query against 2.1 TB of raw packet capture reads 412 KB and returns in 38 milliseconds. At 100 EB, the node count visited is the same. This paper covers the Query Analytics Tree structure, the sub-bitmap AND pruning algorithm, SILO correlation, and the storage math behind indefinite analytical record retention.
Mark Bednarczyk Blog: /blog/qat-architecture April 2026
Download PDF
AI Defense 30 min read
AI Defense Needs a Feature Stream, Not Another Model
Why state management at machine speed starts with the stream, not the model
4 bytes Token Header 800 Gbps Correlation 4 Confidence States 0 Models We Ship
When a beacon flow first crosses the wire, the Protocol Stack emits six tokens — 28 bytes total — before any alert fires. That stream is the substrate AI defense needs. This paper covers the Token Stream format, confidence state transitions, and why feature engineering at 800 Gbps requires the stream to precede the model.
Mark Bednarczyk Blog: /blog/token-stream April 2026

Each paper is the full technical treatment. The companion blog articles are at /blog.

Want to Go Deeper?

We're happy to walk through the architecture, answer technical questions, or tailor a demo to your environment.

Talk to the Team Request a Demo