Global Federated Stable Interoperability Protocol GFSIP v1.0
[简体中文](README-zh.md) | English
ANTARES is GFSIP v1.0 — the Global Federated Stable Interoperability Protocol. It provides encrypted, multiplexed, recoverable cross-domain communication for services, AI agents, devices, and organizations, without a central authority. With QUIC as the mandatory transport layer and deterministic CBOR serialization, it has built-in session resumption, idempotent deduplication, and causal audit, serving as the stable foundation of a decentralized network.
— ✦ —
| # | Capability | Description |
|---|---|---|
| 1 | Encrypted Session | Mutually authenticated TLS or token-authenticated sessions over QUIC |
| 2 | Multiplexing | Independent logical channels within a single session |
| 3 | Session Resumption | Resume sessions without re-authentication after network handover |
| 4 | Idempotent Side Effects | Windowed deduplication via idempotency keys |
| 5 | Causal Audit (optional) | Signed event records with declared causal predecessors |
| 6 | Cross-Domain Federation | Multi-trust-anchor routing + signed domain descriptors |
| 7 | Fault Isolation | A single domain failure does not block intra-domain traffic |
— ✦ —
GFSIP uses a 44-byte fixed-length header over QUIC, with ALPN gfsip/1. Extension headers use deterministic CBOR (big-endian, shortest encoding). 18 standard message types (0x01–0x16) cover the full lifecycle: handshake, channel management, data transfer, session resumption, audit events, and federation routing.
— ✦ —
| File | Purpose |
|---|---|
GFSIP_v1.0_protocol_spec.md | Full protocol specification (33 sections) |
gfsip-state-machine.json | Machine-readable state machine definition |
gfsip-message-schema.json | Logical message JSON Schema |
gfsip-error-registry.json | Numeric error code registry |
gfsip-conformance-checklist.csv | Conformance test checklist |
reference-impl/ | Python reference implementation — frame codec, state machine, channel management, authentication, deduplication, resumption, audit, federation, end-to-end demo (9/9 conformance vectors passing) |
— ✦ —
Core/1 (mandatory) — QUIC transport, version negotiation, mutual authentication, session and channel management, session resumption, structured errors, graceful shutdown.
Audit/1 (optional) — Signed causal event records with declared causal predecessors, actor identity, rule version, and pre/post state hashes; rejects self-loops and known cycles.
Federation/1 (optional) — Cross-domain routing via signed DomainDescriptor objects, multi-trust-anchor configuration, descriptor expiration / revocation, and fault isolation.
— ✦ —
# Install SDK (PyPI)
pip install gfsip
# Run end-to-end demo (clone the repo to get demo files)
git clone https://github.com/nohn3043-arch/Antares.git
cd Antares/reference-impl
pip install -r requirements.txt
python demo.py # Sync API demo: handshake -> channel -> data -> dedup -> resume
python async_demo.py # Async API demo: await connect/open_channel/send (zero manual pump)
python gfsip/conformance.py # 9/9 Section 28 minimal conformance vectors — all passing— ✦ —
- AI Agent Networks — Multi-agent task collaboration across organizational boundaries, with auditable trails
- Enterprise Integration — Cross-organizational workflow orchestration without a central authority
- IoT / Edge — Device-to-cloud session resumption under network handover
- Finance / Compliance — Signed causal audit chains for regulatory reporting
- Healthcare — Federated data exchange with domain-level policy enforcement
- Robotics — Reliable command channels for idempotent safe operations
— ✦ —
| Milestone | Status |
|---|---|
| Spec v1.0 (interface frozen) | ✅ Done |
| Machine-readable state machine and schemas | ✅ Done |
| Python reference implementation (9/9 conformance) | ✅ Done |
| PyPI package release (pip install gfsip) | ✅ Done |
| Async high-level API (AsyncEndpoint, zero manual pump) | ✅ Done |
| Independent second implementation | 🔲 In progress |
| Public interoperability report | 🔲 Todo |
| Production pilot domain | 🔲 Todo |
— ✦ —
ANTARES is a member of the NOHN AI ecosystem — a family of projects built around second-perspective causal audit and deterministic execution:
| Project | Repository | Role |
|---|---|---|
| Second-Perspective (GCAE) | nohn3043-arch/second-perspective | Global cognitive audit engine — five-operator causal audit core (IMDA 95/100) |
| NOMOS | nohn3043-arch/second-perspective (Intelligent-Decision-Hub--Nomos branch) |
Auditable deterministic decision hub (IMDA 95/100) |
| SPL-G1 | nohn3043-arch/SPL-G1 | Hardware causal-audit trusted compute unit (TCU) |
| SPL-Virtual-World-Base | nohn3043-arch/Second-Reality | Virtual-world and metaverse infrastructure (Constitution / Law / Bridge) |
| Story-Engine | nohn3043-arch/story-engine | Long-form narrative consistency engine |
| Antares | nohn3043-arch/Antares | GFSIP v1.0 — federated stable interoperability protocol with causal audit |
| Anthropomorphic-Agent-Engine | nohn3043-arch/Anthropomorphic-Agent-Engine | Deterministic anthropomorphic psychology engine (SPL Pure Core V8.0) |
| PAGES | nohn3043-arch/pages | Official NOHN AI ecosystem landing page |
— ✦ —
This repository uses a dual-track license:
| Content | License | File |
|---|---|---|
Code implementation (the Python package gfsip under reference-impl/) | Apache License 2.0 | LICENSE |
Protocol specification and documentation (GFSIP_v1.0_protocol_spec.md, gfsip-*.json, gfsip-*.csv, assets/ diagrams) | CC BY 4.0 | LICENSE-SPEC.md |
Apache-2.0 permits commercial use, modification, and distribution, and includes an express patent grant; CC BY 4.0 permits any use and redistribution (including commercial use) with attribution only — the open attribution license for the specification serves GFSIP's goal of dissemination as an open standard.
Irrevocable notice: Versions up to and including v1.0 were previously released in their entirety under CC BY 4.0 (including code). That license is irrevocable under its terms, and those versions will remain permanently usable under CC BY 4.0. The dual-track arrangement above takes effect from the next version.
- Contact: International / Global — ai@nohnlins.com · China — lin@secondai.top
GitHub · nohnlins.com · ai@nohnlins.com
NOHN AI · ANTARES