Bunny Database

Product Introduction

1. Definition: Bunny Database is a globally distributed, SQLite-compatible Database-as-a-Service (DBaaS) running on bunny.net’s edge network. It automatically spins down compute resources during inactivity.
2. Core Value Proposition: It eliminates cold starts and latency issues for globally distributed applications by deploying databases in 41 regions, ensuring low-latency access while reducing costs via idle spin-down.

Main Features

1. SQLite Compatibility: Uses libSQL (SQLite fork) for full compatibility with SQLite syntax, tools, and libraries. Supports HTTP API and SDKs (TypeScript/JavaScript, Go, Rust, .NET) for seamless integration.
2. Automatic Spin-Down: Halts compute resources during inactivity, charging only for storage ($0.10/GB/month). Wakes in <100ms upon request via proprietary activity detection.
3. Multi-Region Replication: Deploy read replicas across 41 global regions via dashboard. Async replication syncs from the primary region, reducing latency by up to 99% for distant users.
4. Pay-Per-Use Pricing: Metered billing for reads ($0.30/billion rows), writes ($0.30/million rows), and storage. Replicas incur compute costs only when active (hourly metering).
5. Integrated Observability: Real-time dashboard tracks query latency (p50/p95), storage usage, and traffic. Includes SQL editor for schema inspection and query execution.

Problems Solved

1. Pain Point: High latency for global users due to centralized databases. Traditional solutions (e.g., AWS RDS) require manual replication and incur 24/7 costs.
2. Target Audience:
   • Frontend developers building low-latency Jamstack apps.
   • Startup CTOs needing cost-effective scaling for global user bases.
   • DevOps teams managing read-heavy workloads (e.g., catalogs, user profiles).
3. Use Cases:
   • Product catalogs with region-specific pricing/inventory.
   • User profile storage for social apps with global audiences.
   • Configuration databases for edge-deployed serverless functions.

Unique Advantages

1. Differentiation: Unlike Cloudflare D1 or Turso, Bunny Database leverages bunny.net’s private backbone (1,300+ PoPs) for sub-50ms global reads. Idle spin-down reduces costs by 90%+ vs. always-on competitors.
2. Key Innovation: Deep integration with Bunny Edge Scripting and Magic Containers enables unified deployment of compute + database at the edge, avoiding data-transfer fees.

Frequently Asked Questions (FAQ)

1. How does Bunny Database achieve low-latency global reads?
   By deploying read replicas in 41 regions via bunny.net’s Anycast network, routing queries to the nearest replica.
2. Is data persisted during spin-down?
   Yes. NVMe SSD storage remains active ($0.10/GB/month), while compute halts. Databases auto-wake for queries.
3. Can I migrate existing SQLite databases?
   Yes. Upload .db files via dashboard or connect directly using libSQL-compatible SDKs.
4. What happens if the primary region fails?
   Manual failover required. Bunny recommends pairing with Magic Containers for high-availability workloads.
5. How are writes handled in multi-region setups?
   Writes route to the primary region only, with async replication to read replicas. Use Edge Scripting for write coalescing to minimize latency.