Fabric-X Committer

The Fabric-X Committer implements the validation and commit stage of the Hyperledger Fabric-X transaction lifecycle. It receives ordered blocks of transactions, verifies endorsement signatures, performs MVCC (Multi-Version Concurrency Control) validation against the state database, and commits valid transactions — all through a pipelined architecture designed for high throughput.

Architecture at a Glance

The system is built around six core services connected by gRPC streams and bounded channels:

Ordering Service
        │
        ▼
    Sidecar ──────────► Coordinator
                         │        │
                  Verify │        │  Validate & Commit
                         ▼        ▼
                     Verifier     VC ──► Database Cluster
                                             ▲
                                  Query ─────┘
                                 Service

• Sidecar — Fetches blocks from the Ordering Service, persists them locally, and delivers committed blocks to clients. Exposes a Notification Service for asynchronous transaction status updates.
• Coordinator — Orchestrates the pipeline using a transaction dependency graph to identify and dispatch conflict-free transactions for parallel processing.
• Verifier — Validates transaction signatures against namespace endorsement policies using a parallel worker pool.
• Validator-Committer (VC) — Executes a three-stage pipeline (Prepare, Validate, Commit) performing MVCC checks and committing valid transactions to the database.
• Query Service — Provides read-only access to the committed world state with configurable isolation levels and query batching.
• Database Cluster — Stores world state, transaction statuses, and namespace policies. Supports YugabyteDB (recommended) and PostgreSQL.

Key Capabilities

• High Throughput — Pipelined processing with parallel dispatch of conflict-free transactions. Exceeds 100,000 TPS on commodity hardware with YugabyteDB.
• Fault Tolerance — Idempotent commit operations enable automatic recovery from service failures without data corruption. Each service recovers independently on restart.
• Horizontal Scaling — Verifier, VC, Query Service, and Database nodes scale horizontally. Sidecar and Coordinator scale vertically.
• Flexible Endorsement Policies — Supports both lightweight threshold rules (single public key) and fine-grained MSP rules (AND/OR/k-of-n over organizational identities).
• Observability — Prometheus metrics for every pipeline stage, with queue-depth gauges for bottleneck identification.

Getting Started

See the Setup Guide for prerequisites, build instructions, and running tests.

For production deployments, start with the Deployment Guide for hardware sizing and topology, then the Performance Tuning guide for configuration parameters.

Documentation Overview

Section	What You'll Find
Architecture	System design, component interactions, state management, failure recovery
Deployment	Hardware requirements, reference topology, scaling guidance, startup order
Services	Detailed documentation for each service: workflows, APIs, configuration, recovery
Configuration	TLS setup, logging, metrics reference, namespace policies
Development	Internal patterns (errgroup, context-aware channels) and load generator tooling