Backend

Pluggable storage abstraction layer supporting different storage implementations.

Architecture

The backend system has two layers:

• BackendImpl trait: The storage trait that backends implement
• Backend wrapper: Instance-level wrapper providing future local/remote dispatch

BackendImpl Trait

Abstracts underlying storage to allow different backends without changing core logic.

Core Operations:

• Entry storage and retrieval by content-addressable ID
• Verification status tracking for authentication
• Database and store tip calculation
• Topological sorting for consistent entry ordering

Current Implementation

InMemory: HashMap-based storage with JSON file persistence

• Stores entries and verification status
• Includes save/load functionality for state preservation
• Supports all BackendImpl trait operations

Verification Status

Verified: Entry cryptographically verified and authorized

Unverified: Entry lacks authentication or failed verification

Status determined during commit based on signature validation and permission checking.

Key Features

Entry Storage: Immutable entries with content-addressable IDs

Tip Calculation: Identifies entries with no children in databases/stores

Height Calculation: Computes topological heights for proper ordering

Graph Traversal: Efficient DAG navigation for database operations

Custom Backend Implementation

Implement BackendImpl trait with:

1. Storage-specific logic for all trait methods
2. Verification status tracking support
3. Thread safety (Send + Sync + Any)
4. Performance considerations for graph operations

The Backend wrapper will automatically delegate operations to your BackendImpl implementation.