It's now available:

PON-DB
The Solid State
Database Engine.

O(1) lookup on AES-256-GCM encrypted, compressed datasets.
Zero tuning. Single static binary.

~0.08ms

Avg query latency

0.1%

CPU per query

~0.1MB

RAM per query

100M+

rows tested

The Problem

Historically, all database engines have forced you to choose.

⚡ Speed vs Security

Fast databases skip encryption. Encrypted databases add latency layers, key management overhead, and external dependencies. You compromise one for the other.

⚙ Performance vs Simplicity

High-performance engines need servers, daemons, tuning, DBA expertise, and ongoing maintenance. Simple solutions don't scale. Nothing in between exists.

📦 Portability vs Capability

Embedded databases are limited. Powerful databases require infrastructure. Deploying a capable database to edge, IoT, or air-gapped environments is an unsolved problem.

PON-DB The Solid State Database Engine. doesn't.


Features

Built different. By design.

Seek O(1)

Performance and Scalability Constant-time lookup Algorithm

RA-CHAMPI algorithm resolves key-value reads in O(1) regardless of dataset size. Tested at 100M+ rows with no latency degradation.

  • Implementation of Seek O(1) index lookup in read-only mode, validated on datasets with over 100 million rows.
  • Deterministic Indexing: Generation of a single key field per ingestion to guarantee direct and predictable access.
  • Ultra-Low Latency: Sub-millisecond performance on cold starts and under high-concurrency scenarios.
  • Zero Query Tuning: Self-optimizing engine that eliminates the need for manual index configuration or database tuning.
  • Resource Efficiency: Extremely low consumption (<1MB RAM and ~0.1% CPU per query), optimized using Rust Zero-Cost Abstractions.
  • Operational Immutability: Optimized processing of the original JSON using a post-ingestion read-only state for maximum stability.
  • Auto-optimization based on CPU/RAM resources. :Dynamic adjustment of search and cache algorithms based on the availability of real-time CPU and RAM telemetry.
AES 256 GCM

Data Security and Protection Military-Grade Encryption

Encryption at rest and in transit, authenticated. No plugins, no external key managers. Every .pon file is encrypted by default.

  • Native implementation of AES-256-GCM (Authenticated Encryption), guaranteeing the confidentiality and integrity of data both at rest and in transit.
  • Secure Memory Management (Data Zeroization): Active and guaranteed cleaning of sensitive data in RAM after each operation, mitigating memory inspection or cold boot attacks.
  • Cryptographic Integrity: Storage designed using cryptographically verifiable structures, allowing data authenticity to be audited at any time.
  • Architectural Immunity: Robust design against common attacks; by not relying on text interpretation engines, the system is inherently immune to SQL injections and buffer overflows.
Byte-Exact

Data Integrity and Accuracy High-Fidelity Deterministic Engine

  • Architecture designed to preserve the absolute integrity of the original JSON bytes during processing.
  • Byte-Exact Reversibility: Guarantee of identical reconstruction of the original data, backed by integrity validation using SHA-256.
  • Total Flow Transparency: Absolute consistency between input and output; the software guarantees no mutations, loss of numerical precision, or schema changes during the query.
  • Zero-Copy Deserialization Efficiency: Implementation of intelligent pointers to access data without copying it into memory, maximizing speed and preserving the original structure.
Portability

Portability and Deployment, Single Static Binary

  • Single, standalone executable compiled for x86-64 architecture (~1 MB).
  • Zero-Dependency Architecture: Native operation without requiring external libraries, runtimes (such as Java or Node.js), or operating system dependencies.
Compression

Adaptive Compression, Entropy-Based Compression

Size reduction algorithms that dynamically adapt according to the density and repetitiveness of the JSON structure, optimizing storage without sacrificing decompression speed.

Interoperability

Universal Integration via CLI/JSON (Interoperability)

Each .pon file can be bound to specific physical hardware. Data that physically cannot run anywhere else. Built-in DRM at the storage layer.

  • Polyglot Interface: Ability to perform queries and filters using system calls in multiple languages ​​(Python, C, Rust, JavaScript, PHP, Go, and others).
  • Structured Standard Output: Direct communication via STDOUT in JSON format, enabling immediate and agnostic integration across development ecosystems.

More options

Custom options, for corporate licenses or special orders:


Benchmark

Same hardware. Xeon 4C 24Gb/RAM
Different reality.

PON-DB v8.4, test per query in CLI
  • RAM usage 0.1 MB
  • CPU usage 1 - 3 %
  • Query latency 0.10 ms average
  • Dataset:NY Taxi dataset 2023 20.2GB / 38310226 rows
  • Encryption AES-256-GCM ✓
M****DB (same server)
  • RAM usage 10.6 GB
  • CPU usage 19.3%
  • Query latency varies
  • Dataset standard
  • Encryption plugin required

Pricing

Permanent license per physical machine Forever.

Unlimited Instances.

Standard Early Adopters,only 100 Licenses!

$799

one-time · per machine


  • Seek O(1) Key Value Read Mode Constant-time lookup(see our bechmarks)
  • Native AES-256-GCM encryption by row
  • Deterministic Indexing,a single key field per ingestion to guarantee direct and predictable access
  • Zero Query Tuning
  • Architectural Immunity to SQL injections and buffer overflows
  • Secure Memory Management (Data Zeroization)
  • Extremely low consumption using Rust Zero-Cost Abstractions
  • Single, standalone executable compiled for x86-64 architecture (~1 MB)
  • Zero-Dependency & Zero Trust Architecture
  • Index Field Entropy-Based Compression
  • Up to 10M rows per .pon file
  • Unlimited instances in the same machine
  • Universal Integration via CLI/JSON
  • Support by email

only few licenses remain, act now!
after that the regular price will be $1,598

Enterprise / Corporate

Contact

custom · multi-node


  • Seek O(1) Key Value Read Mode Constant-time lookup (see our bechmarks)
  • Native AES-256-GCM encryption by row
  • Deterministic Indexing,a single key field per ingestion to guarantee direct and predictable access
  • Zero Query Tuning
  • Architectural Immunity to SQL injections and buffer overflows
  • Secure Memory Management (Data Zeroization)
  • Extremely low consumption using Rust Zero-Cost Abstractions
  • Single, standalone executable compiled for x86-64 architecture (~1 MB)
  • Zero-Dependency & Zero Trust Architecture
  • Index Field Entropy-Based Compression
  • 100M+ rows per .pon file
  • Unlimited instances in the same machine
  • Universal Integration via CLI/JSON
  • --self option: Cryptographic binding of the data file to the physical hardware (CPU/Disk ID) to prevent unauthorized portability of the database.
  • Multi-core processing: the software detects how many cores the machine has and adapts performance to take advantage of all available cores.
  • Dedicated & Custom integration support

Limits enforced by the engine itself—not by policy. The binary physically stops processing beyond your licensed tier; for example, if you try to process 100M rows with a 10M rows license, you'll only get 10M rows. The limit is based on the number of rows in the JSON file you want to process. You can process as many JSON files as you want.