Network Topology

The structure and organization of the ComputeNet network

Draft Documentation

This documentation is under development and may be incomplete or subject to change.

Network Structure

ComputeNet uses a decentralized peer-to-peer topology where nodes communicate directly without central coordination. The network self-organizes into overlapping validator sets for efficient job distribution.

Node Types

Validator Nodes

Execute compute jobs, generate proofs, and participate in consensus. Require staking and meet hardware requirements.

Full Nodes

Store complete history and verify all proofs. Do not execute jobs but provide data availability.

Light Clients

Verify specific receipts without storing full history. Suitable for end-user applications.

Gateway Nodes

Provide RPC interfaces for external applications. Bridge between clients and the network.

Peer Discovery

Nodes discover peers through a combination of bootstrap nodes, DHT-based discovery, and gossip protocols.

# Example node configuration
[network]
bootstrap_peers = [
  "/dns4/boot1.computenet.io/tcp/30333",
  "/dns4/boot2.computenet.io/tcp/30333"
]
max_peers = 50
enable_dht = true

Geographic Distribution

For optimal performance and fault tolerance, validators are encouraged to distribute geographically. The protocol includes latency-aware job routing to minimize execution time while maintaining verification guarantees.

PreviousVerification ProtocolNextConsensus