Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Database

Represents an independent, versioned collection of data entries within Eidetica, analogous to a database in traditional databases.

Conceptual Model

Databases organize related data entries into a coherent unit with its own history and authentication policies. Each Database is identified by its root entry's content-addressable ID, making it globally unique and verifiable.

Unlike traditional databases, Databases maintain full historical data through a Merkle DAG structure, enabling features like:

  • Conflict-free merging of concurrent changes
  • Cryptographic verification of data integrity
  • Decentralized synchronization across devices
  • Point-in-time queries (unimplemented)

Architecture and Lifecycle

Database Creation: Initialized with settings (stored as a Doc CRDT) and associated with an authentication key for signing operations.

Data Access: Applications interact with Databases through Transaction instances, which provide transactional semantics and store access.

Entry History: Each operation creates new entries that reference their parents, building an immutable history DAG.

Settings Management: Database-level configuration (permissions, sync settings, etc.) is stored as CRDT data, allowing distributed updates.

Authentication

Each Database maintains its own authentication configuration in the special _settings store. All entries must be cryptographically signed with Ed25519 signatures - there are no unsigned entries in Eidetica.

Databases support direct keys, delegation to other databases for flexible cross-project authentication, and a three-tier permission hierarchy (Admin, Write, Read) with priority-based key management. Authentication changes merge deterministically using Last-Write-Wins semantics.

For complete details, see Authentication.

Integration Points

Store Access: Databases provide typed access to different data structures (DocStore, Table, YDoc) through the store system.

Synchronization: Databases serve as the primary unit of synchronization, with independent merge and conflict resolution.