Internal

Technical architecture and implementation details for MuduDB internals: kernel design, adapter model, partitioning, worker model, and the procedure runtime. Intended for developers and contributors.

• Kernel
  • Kernel-runtime split
  • Responsibilities
  • Worker model
  • Boundaries and compatibility
• Adapters
  • Supported backends
  • Connection configuration
  • Async session loop
  • Known differences
• Procedure Runtime
  • Design
  • Synchronous source, asynchronous execution
  • Syscall boundary
  • Use cases
• WebAssembly Integration
  • Component model
  • WASI sandbox
  • Language support
  • Guest bindings
  • Runtime limits
  • Package format
• Partitioning and Worker Mapping
  • Overview
  • Design rationale
  • Terminology
  • Model & syntax
  • How routing works
  • Discovering topology
  • Backend mapping & caching
  • Sharding port and worker connection
  • Obtaining topology and connecting directly
  • Usage examples
  • Failure modes & fallbacks
  • Operational concerns
  • Performance and tuning
  • Troubleshooting
• Per-core Worker Model
  • 1. Overview
  • 2. Motivation
  • 3. Per-worker model
  • 6. Inter-worker communication
  • 7. Data locality and partitioning