1.9 KiB

Raw Blame History

idea134-matchops-federated-verifiable

MatchOps is a federated session-orchestration toolkit for multiplayer games. It models regions, servers, session types, capacity budgets, latency budgets, fairness rules, and execution policies in a declarative DSL, then compiles that spec into canonical IR, a verifiable execution graph, and governance artifacts.

This repository focuses on the control-plane primitives:

declarative DSL compilation
canonical IR and plan deltas
cryptographically hashed execution attestations
a Graph-of-Contracts adapter registry
a privacy-preserving federated scheduler
a tamper-evident governance ledger backed by SQLite

What It Provides

MatchOpsSpec: the input DSL model.
compile_spec(...): converts a spec into canonical IR and a verifiable execution graph.
FederatedOptimizer: allocates session demand from aggregated regional signals without exposing raw player data.
GoCRegistry: registers versioned adapters for cloud APIs and game-server routing.
SQLiteGovernanceLedger: stores append-only governance events with a hash chain.

Repository Layout

src/idea134_matchops_federated_verifiable/
tests/
AGENTS.md
test.sh
pyproject.toml

Quick Start

python3 -m pip install -e .
pytest -q
python3 -m build

Or run the full verification script:

bash test.sh

Design Notes

Canonical JSON serialization is used for all hashes so decisions are reproducible.
The execution graph uses per-step attestations derived from the previous step and the current payload.
The optimizer only consumes aggregated regional signals and session demand, never raw player identity data.
The governance ledger is append-only and can be persisted to SQLite for durable audit trails.

Public Package Metadata

The package is published as idea134-matchops-federated-verifiable and its README is wired into pyproject.toml via readme = "README.md".

1.9 KiB Raw Blame History