Native storage and recovery.
8 KB pages, buffer pool, checksums, WAL and automatic crash recovery; built independently in Rust.
Engineered in Mexico · Built for what matters
Built from scratch. Running for real.
A functional relational SQL database engine built from scratch in Rust. Native storage, MVCC transactions, SQL processing and enterprise administration.
Advanced technical alpha · Functional SQL engine · Not certified for critical production
piya / sql-previewTECHNICAL ALPHA
-- Illustrative SQL · not a live engine
BEGIN;
SELECT customer_id,
SUM(amount) AS total
FROM transactions
WHERE status = 'settled'
GROUP BY customer_id
ORDER BY total DESC;
COMMIT;Illustrative SQL · Alpha engine availableSQL / Rust
01 / Project status
Functional core. Enterprise maturity in progress.
PiyaDB has moved beyond conceptual architecture: a working SQL database engine exists today. Here is what the technical alpha includes and what still requires validation.
ACID + MVCC
READ COMMITTED, snapshot-based isolation, savepoints, rollback and referential integrity.
SQL that executes.
Joins, subqueries, analytical functions, DDL/DML, views, hash indexes, sequences, synonyms and SQL-compatible built-in functions.
Drivers and SB-Wire.
Native binary protocol and client-server foundation exist. Drivers require maturity and compatibility consolidation.
Reliability at enterprise scale.
Persistent B+Tree indexes, advanced optimization, online/incremental backups, PITR, replication, HA, failover, procedural SQL, encryption hardening and long-duration stress tests.
Technical alpha is not a production certification. Implementation status does not imply validated SLAs, universal SQL compatibility or critical workload readiness.
01 / Foundation
Built from first principles.
A modular database architecture with its own SQL execution, transaction, and persistent storage layers—designed without depending on another RDBMS for data persistence.
Independent by architecture.
Implemented SQL execution, query optimization, statistics, execution plans, catalog and native storage engine.
CLI / DRIVERS
SQL PROTOCOL
MANAGEMENT API
SQL ENGINE
ACID / MVCC
RBAC / TLS
BUFFER POOL
WAL / RECOVERY
PAGE STORAGE
Rust at the core.
Rust-native modular architecture with installation and execution mechanisms for Linux and macOS.
Transactions that matter.
Implemented MVCC transactions, READ COMMITTED, snapshot isolation, savepoints, rollback and referential integrity.
Resilience by design.
Roadmap for physical and logical replication, failover, fencing, backups, and point-in-time recovery.
Security as a foundation.
Users, roles, privileges and auditing are implemented. Encryption and advanced hardening remain in development.
02 / Operations
The engine and the control plane. Together.
A React and TypeScript enterprise console already provides SQL workspace, users, roles, object exploration, sessions, performance monitoring, auditing and diagnostics.
Follow the roadmap ↗ENTERPRISE CONSOLE / ALPHA MODULESFUNCTIONAL ALPHA
- 01Dashboard / Health & Telemetry
- 02SQL Workspace / Execution Plans
- 03Performance / Sessions / Storage
- 04Users / Roles / Privileges / Audit
- 05Backup / HA — in development
03 / Ecosystem
Designed to connect.
PiyaDB includes a client-server architecture, CLI, management API and its own SB-Wire binary protocol. Driver maturity varies; examples below are conceptual, not published SDK contracts.
rust / conceptual integration
Conceptual pseudocode — API not finalizedSDK ROADMAP
04 / Engineering validation
Evidence before claims.
Unit, integration, concurrency, recovery, security and SQL compatibility tests exist; the latest documented run reports zero failures. Reproducible scale, soak and performance benchmarks remain pending.
| Metric | Validation approach | Published result |
|---|---|---|
| Transaction throughput (TPS) | Repeatable transactional workloads | Pending validation |
| Query latency (p50 / p95 / p99) | Controlled SQL workloads | Pending validation |
| Memory & CPU utilization | Documented hardware profiles | Pending validation |
| Recovery / RPO / RTO | Failure injection & recovery testing | Pending validation |
| SQL interoperability | Published functional compatibility matrix | Pending validation |
No invented benchmark numbers, unsupported superiority claims, or implied production certifications. Architecture goals are not performance guarantees.
05 / Technical FAQ
The important questions.
Is PiyaDB production-ready?
No. PiyaDB already has a functional SQL engine in advanced technical alpha. Large-scale, long-running stability and critical production suitability still require validation.
How does PiyaDB approach SQL compatibility?
PiyaDB is an independent relational SQL engine. Broader SQL dialect compatibility, procedural SQL and client interoperability are evolving capabilities; no universal compatibility or certified migration path is claimed.
What platforms will be supported?
Installation and execution mechanisms exist for Linux and macOS. Official platform, version and support certification matrices are not yet published.
What is the licensing model?
PiyaDB is being positioned as a commercial enterprise product. Specific license terms, pricing, support levels, and availability have not yet been published.
How can I join early access?
You can express interest in receiving development updates and future evaluation opportunities. This is not an active beta or a commitment to product availability.
Be part of what comes next
Follow the next chapter in enterprise data.
Register your interest for project updates and potential future technical evaluations. No beta access is implied.
Express interest ↗
Opens your email client. A registration backend is not configured.