Comparisons
vs plain JSON files#
| Traditional JSON files | AxioDB |
|---|---|
| Single JSON file — heavy read/write I/O | File-per-document storage |
| No built-in caching | InMemoryCache strategy |
| Linear search for document retrieval | Auto-indexed documentId for fast lookups |
| Performance degrades with dataset size | Maintains performance at scale |
| No indexing system | Multi-core parallelism with Worker Threads |
| Manual file locking for concurrency | Built-in query operators ($gt, $lt, $regex, $in) |
vs SQLite#
| Feature | SQLite | AxioDB |
|---|---|---|
| Native dependencies | Yes (C bindings) | Pure JavaScript |
| Query language | SQL strings | JavaScript objects |
| Schema migrations | Required (ALTER TABLE) |
Schema-less (optional) |
| Built-in caching | Manual | InMemoryCache |
| Multi-core processing | Single-threaded | Worker Threads |
| Built-in GUI | External tools only | Web interface included |
| Best for | 10M+ records, relational data | 10K-500K documents, embedded apps |
Document retrieval speed#
AxioDB's file-per-document architecture with InMemoryCache gives instant document retrieval by ID — O(1) with cache, versus a linear O(n) scan for plain JSON files. Unlike SQLite, there's no native dependency to compile or ship.
When AxioDB is not the right choice#
- 10M+ documents, or datasets that need to scale past a single node — use PostgreSQL, MongoDB, or SQLite.
- Multi-user web applications with hundreds of concurrent connections — AxioDB is single-instance, not client-server (though AxioDBCloud adds remote TCP access for service-to-service use).
- Relational data with JOINs and foreign-key constraints — AxioDB is document-based NoSQL.
- Distributed systems needing replication, sharding, or clustering — AxioDB is single-node only.
- Cross-collection ACID transactions — AxioDB's transactions are scoped to a single collection.
AxioDB isn't competing with PostgreSQL or MongoDB — it's for when you need a database embedded in your app, and migrating out when you outgrow it is the expected path, not a failure.