protocol-acset

Protocol ACSet: Compositional P2P Protocol Design

Model decentralized and P2P protocols as attributed C-sets (categorical data structures) to enable compositional analysis, verify interoperability, and design protocol evolution narratives.

Core Insight

Rather than viewing protocols as isolated systems, Protocol ACSet treats them as compositional objects in a category where:

• Objects = Protocols (IPFS, Iroh, Matrix, Nostr, etc.)
• Morphisms = Protocol bridges and adapters
• Attributes = Protocol properties (transport, encryption, topology)
• Composition = How protocols stack and interoperate

What is an Attributed C-Set?

An Attributed C-set is a graph structure with:

• Vertices (objects) and edges (relationships)
• Attributes (data attached to vertices/edges)
• Functorial structure (composition preserving operations)

Example: IPFS as an ACSet

objects:
  - Object(name="ipfs", type="content-distribution", transport="tcp/quic")

morphisms:
  - Morphism(from="ipfs-node", to="ipfs-peer", label="connect")
  - Morphism(from="content-hash", to="blob", label="address")

attributes:
  - (ipfs): [encryption="aes", topology="dht", consensus="none"]
  - (ipfs→peer): [latency=50ms, bandwidth=100mbps]

Protocol Categories (Objects)

Every protocol maps to one or more protocol categories:

ProtocolACSet = {
  objects: {
    transport,          // TCP, QUIC, UDP, WebRTC
    security,           // TLS, Noise, WireGuard
    topology,           // P2P, federated, hybrid
    identity,           // Public keys, DIDs, domains
    data_model,         // Append-only, CRDT, graph
    discovery,          // DHT, mDNS, relay, centralized
    incentive           // Proof-of-work, Filecoin, none
  }
}

Key Protocols as ACSet Objects

Transport Layer

TRANSPORT = {
  objects: [TCP, QUIC, UDP, WebSocket, WebRTC],
  morphisms: {
    TCP.upgrade_to(QUIC),      // 0-RTT, connection migration
    UDP.extend_to(QUIC),       // Congestion control, ordering
    WebSocket.fallback_to(TCP) // Browser compatibility
  },
  attributes: {
    latency: latency,
    connection_setup: rtt_count,
    encryption_native: bool
  }
}

Security Layer

SECURITY = {
  objects: [TLS, Noise, WireGuard, Signal_Double_Ratchet],
  morphisms: {
    Noise.adapt_to(libp2p),    // Generic handshake
    WireGuard.operate_at(layer2),  // VPN vs app layer
    Signal.strengthen_with(perfect_forward_secrecy)
  },
  attributes: {
    forward_secrecy: bool,
    post_quantum_resistant: bool,
    overhead_bytes: int
  }
}

Topology Layer

TOPOLOGY = {
  objects: [P2P_Direct, Federated, Relay_Based, Hybrid],
  morphisms: {
    P2P_Direct.fallback_to(Relay),  // Firewall bypass
    Federated.include_P2P_option,   // FEP-1024 pattern
    Relay.optimize_with_DHT         // Peer discovery
  },
  attributes: {
    centralization_risk: 0..1,
    scalability: metric,
    latency_vs_centralization: tradeoff
  }
}

Identity Layer

IDENTITY = {
  objects: [PublicKey, DID, Domain, Account],
  morphisms: {
    PublicKey.trustless(),         // No registration
    DID.composable_with(VerificationMethod),
    Domain.delegate_to(Protocol),  // FQDN ownership
    Account.require_centralized_server()
  },
  attributes: {
    portability: bool,
    user_controlled: bool,
    recovery_possible: bool
  }
}

Data Model Layer

DATA_MODEL = {
  objects: [AppendOnly, CRDT, GraphDB, RDF, MerkleDAG],
  morphisms: {
    AppendOnly.compose_with(CRDT),     // Git + concurrent edits
    CRDT.guarantee(CommutativeMonoid), // Math foundation
    MerkleDAG.content_address(),       // IPFS pattern
    GraphDB.query_with(SPARQL)
  },
  attributes: {
    eventual_consistency: bool,
    conflict_resolution: strategy,
    query_capability: query_language
  }
}

Composition Laws

Protocols compose along morphisms following categorical laws:

Associativity (Functorial Composition)

(IPFS → IPNS → ENS) = IPFS → (IPNS → ENS)

Both paths represent content resolution chains—they should give the same result.

Identity (No-op Morphism)

Protocol ⊗ Identity = Protocol

Example: A protocol using transparent relays should function identically.

Yoneda Lemma (Composition Universality)

A protocol's properties are determined by how it relates to all other protocols:

IPFS_properties = ∀P: morphisms(P → IPFS) × morphisms(IPFS → P)

If two protocols have identical composition patterns with all others, they're equivalent.

Real Protocols as ACSet Instances

IPFS (Content Distribution)

# ACSet Definition
@acset_type IPFSProtocol <: AbstractACSet begin
  (Node, Content, Peer)
  node_has_content::EdgeVertexDouble
  node_connects_peer::EdgeVertexDouble
  hash_of::EdgeVertex
end

# Instance
ipfs = IPFSProtocol()
add_node!(ipfs, name="node-1", storage=1_000_000)
add_content!(ipfs, hash="baf...", size=50_000)
add_morphism!(ipfs, :node_has_content, 1, 1)  # Node stores content

Iroh (P2P Transport + Blobs)

@acset_type IrohProtocol <: AbstractACSet begin
  (Endpoint, Peer, Blob, Document)
  endpoint_dials::EdgeVertex
  peer_stores_blob::EdgeVertex
  document_syncs::EdgeVertex
  has_relay_fallback::EdgeAttribute
end

iroh = IrohProtocol()
add_endpoint!(iroh, node_id="node-abc", relay_enabled=true)
add_peer!(iroh, peer_id="peer-xyz")
add_blob!(iroh, hash="quic-native")
add_morphism!(iroh, :endpoint_dials, 1, 1)  # Hole punching

Matrix (Federated Messaging)

@acset_type MatrixProtocol <: AbstractACSet begin
  (HomeServer, User, Room, Message)
  user_on_server::EdgeVertex
  message_in_room::EdgeVertex
  server_federates::EdgeVertex
  e2e_encrypted::EdgeAttribute
end

matrix = MatrixProtocol()
add_server!(matrix, domain="matrix.org", e2e=true)
add_user!(matrix, id="@alice:matrix.org")
add_morphism!(matrix, :server_federates, homeserver_alice, homeserver_bob)

Protocol Bridges (Morphisms)

Connect incompatible protocols through adapters:

Bridge: IPFS → ActivityPub (FEP-1024 Pattern)

function ipfs_to_activitypub_bridge(ipfs::IPFSProtocol, ap::ActivityPubProtocol)
  # Map IPFS hash to ActivityPub URL
  bridge = ACSetMorphism()

  for content in ipfs.contents
    url = "ap://link?ipfs=$(content.hash)"
    add_morphism!(bridge, :ipfs_content_to_ap_url, content, url)
  end

  return bridge
end

Bridge: Nostr → Matrix (Relay Adaptation)

function nostr_to_matrix_bridge(nostr::NostrProtocol, matrix::MatrixProtocol)
  # Nostr uses relays, Matrix uses homeservers
  # Adapt relay infrastructure to room subscriptions

  bridge = ACSetMorphism()

  for relay in nostr.relays
    room = add_room!(matrix, name="relay-$(relay.url)")
    add_morphism!(bridge, :relay_to_room, relay, room)
  end

  return bridge
end

Composing Multiple Protocols (Stacking)

Protocol stacking: Use ActivityPub + IPFS + Hypercore simultaneously

# Define the stack
protocol_stack = [
  ActivityPub(),  # Social/federation layer
  IPFS(),        # Content distribution layer
  Hypercore()    # Offline-first data layer
]

# Create composition morphisms
for i in 1:length(protocol_stack)-1
  compose!(protocol_stack[i], protocol_stack[i+1])
  # Each layer morphs through adapters
end

# Result: Hybrid protocol with benefits of all three

Properties preserved through composition:

• ✅ Decentralization (ActivityPub)
• ✅ Permanent links (IPFS)
• ✅ Offline-first sync (Hypercore)

Verification: Interoperability Testing

Use ACSet structure to verify protocols can compose without conflicts:

function can_interoperate(p1::Protocol, p2::Protocol)
  # Check functor compatibility
  return (
    compatible_transport(p1, p2) &&
    compatible_security(p1, p2) &&
    compatible_topology(p1, p2) &&
    compatible_identity(p1, p2) &&
    morphisms_are_natural(p1, p2)
  )
end

# Example
can_interoperate(Nostr(), Matrix())  # true (both message-based)
can_interoperate(IPFS(), Nostr())    # false (content vs events)
can_interoperate(IPFS(), ActivityPub())  # true (FEP-1024)

Protocol Evolution Narratives (Bumpus Sheaves)

Apply Bumpus sheaves on time categories to understand protocol evolution:

Time ──────────────────────────────
  t=2010: Bitcoin (proof-of-work)
    ↓ (evolves with new features)
  t=2014: Ethereum (smart contracts)
    ↓ (layer 2 solutions emerge)
  t=2019: IPFS (content distribution)
    ↓ (bridges to social networks)
  t=2024: ActivityPub + IPFS (hybrid)
    ↓ (emerging toward full P2P)
  t=2025: Multi-protocol stacks

Each time interval in this narrative is a categorical sheaf that captures the state of all protocols at that moment and their mutual morphisms.

GF(3) Balance in Protocols

Assign GF(3) trits to classify protocol design philosophy:

Protocol	Type	Trit	Role
IPFS	Content distrib	+1	PLUS (generative, permanent)
Matrix	Messaging	0	ERGODIC (balanced, federated)
Secure Scuttlebutt	Social	−1	MINUS (observational, offline-first)
Nostr	Social relays	0	ERGODIC (balanced observation/participation)
Iroh	Transport	+1	PLUS (enables new P2P applications)

Conservation law: Sum of trits in a protocol stack = 0 (mod 3)

Example composition: IPFS(+1) ⊗ Matrix(0) ⊗ SSB(−1) = 0 ✓

Building Applications with Protocol ACSet

Step 1: Model Your Application

app_acset = ProtocolStackACSet(
  layers=[
    Protocol(:data, :crdt, :hypercore),
    Protocol(:identity, :public_key, :portable),
    Protocol(:discovery, :dht, :p2p),
    Protocol(:transport, :quic, :encrypted)
  ]
)

Step 2: Verify Composition

verify_composition(app_acset)  # Checks trit balance, morphism compatibility

Step 3: Implement Stack

# Map ACSet to code
for protocol in app_acset.layers
  import_protocol(protocol)
  bind_to_framework(protocol)
end

app = build_from_acset(app_acset)

Use Cases

Use Case	Protocol Stack
Decentralized Social	ActivityPub + IPFS + Hypercore
Offline-First Collab	CRDT (Yjs) + Hypercore + Holepunch
Content Archive	IPFS + IPNS + Filecoin
Private Messaging	Signal (E2E) + Matrix (Federation) + Tor (Privacy)
File Sync	Syncthing (P2P) + IPFS (Distribution) + Ceph (Backup)
Data Sovereignty App	Iroh (Transport) + IPFS (Storage) + DIDs (Identity)

Mathematical Foundations

Functorial Laws

Protocols respect functorial composition:

Protocols → Categories
  F: Protocol → Category
  G: Category → Properties

G(F(IPFS)) = Properties(IPFS)
G(F(Matrix)) = Properties(Matrix)

Natural Transformations

When protocols bridge, they form natural transformations:

IPFS ──bridge──> ActivityPub
  │                  │
  v (F)              v (G)
Properties    →    Properties

The bridge is "natural" if it preserves all protocol properties without loss.

Resources for Deep Dive

• Category Theory: MacLane, "Categories for the Working Mathematician"
• Sheaves on Time Categories: Bumpus, et al. papers
• ACSet Theory: Evan Patterson's work at Topos Institute
• Protocol Design: IETF specifications, protocol RFCs

Conclusion

Protocol ACSet enables:

• ✅ Formal composition of protocols without ad-hoc integration
• ✅ Verification that protocols can safely interoperate
• ✅ Evolution narratives showing protocol families over time
• ✅ Optimal stacking guided by categorical mathematics
• ✅ Data sovereignty through principled system design

Rather than asking "which protocol should I use?", ask "what properties do I need, and how do I compose protocols to achieve them?"

Scientific Skill Interleaving

This skill connects to the K-Dense-AI/claude-scientific-skills ecosystem:

Annotated Data

• anndata [○] via bicomodule
  • Hub for annotated matrices

Bibliography References

• general: 734 citations in bib.duckdb

Cat# Integration

This skill maps to Cat# = Comod(P) as a bicomodule in the equipment structure:

Trit: 0 (ERGODIC)
Home: Prof
Poly Op: ⊗
Kan Role: Adj
Color: #26D826

GF(3) Naturality

The skill participates in triads satisfying:

(-1) + (0) + (+1) ≡ 0 (mod 3)

This ensures compositional coherence in the Cat# equipment structure.