datahaven

DataHaven

AI-First Decentralized Storage secured by EigenLayer — a verifiable storage network for AI training data, machine learning models, and Web3 applications.

Overview

DataHaven is a decentralized storage and retrieval network designed for applications that need verifiable, production-scale data storage. Built on StorageHub and secured by EigenLayer's restaking protocol, DataHaven separates storage from verification: providers store data off-chain while cryptographic commitments are anchored on-chain for tamper-evident verification.

Core Capabilities:

• Verifiable Storage: Files are chunked, hashed into Merkle trees, and committed on-chain — enabling cryptographic proof that data hasn't been tampered with
• Provider Network: Main Storage Providers (MSPs) serve data with competitive offerings, while Backup Storage Providers (BSPs) ensure redundancy through decentralized replication with on-chain slashing for failed proof challenges
• EigenLayer Security: Validator set secured by Ethereum restaking — DataHaven validators register as EigenLayer operators with slashing for misbehavior
• EVM Compatibility: Full Ethereum support via Frontier pallets for smart contracts and familiar Web3 tooling
• Cross-chain Bridge: Native, trustless bridging with Ethereum via Snowbridge for tokens and messages

Architecture

DataHaven combines EigenLayer's shared security with StorageHub's decentralized storage infrastructure:

…

How Storage Works

1. Upload: User selects an MSP, creates a bucket, and uploads files. Files are chunked (8KB default), hashed into Merkle trees, and the root is anchored on-chain.
2. Replication: The MSP coordinates with BSPs to replicate data across the network based on the bucket's replication policy.
3. Retrieval: MSP returns files with Merkle proofs that users verify against on-chain commitments.
4. Verification: BSPs face periodic proof challenges — failure to prove data custody results in on-chain slashing via StorageHub pallets.

Repository Structure

…

Each directory contains its own README with detailed information. See:

• contracts/README.md - Smart contract development
• operator/README.md - Node building and runtime development
• test/README.md - E2E testing and network deployment
• deploy/README.md - Kubernetes deployment
• tools/README.md - Development tools

Quick Start

Prerequisites

• Kurtosis - Network orchestration
• Bun v1.3.2+ - TypeScript runtime
• Docker - Container management
• Foundry - Solidity toolkit
• Rust - For building the operator
• Helm - Kubernetes deployments (optional)
• Zig - For macOS cross-compilation (macOS only)

Launch Local Network

The fastest way to get started is with the interactive CLI:

cd test
bun i                    # Install dependencies
bun cli launch           # Interactive launcher with prompts

This deploys a complete environment including:

• Ethereum network: 2x EL clients (reth), 2x CL clients (lodestar)
• Block explorers: Blockscout (optional), Dora consensus explorer
• DataHaven node: Single validator with fast block times
• Storage providers: MSP and BSP nodes for decentralized storage
• AVS contracts: Deployed and configured on Ethereum
• Snowbridge relayers: Bidirectional message passing

For more options and detailed instructions, see the test README.

Run Tests

cd test
bun test:e2e              # Run all integration tests
bun test:e2e:parallel     # Run with limited concurrency

NOTES: Adding the environment variable INJECT_CONTRACTS=true will inject the contracts when starting the tests to speed up setup.

Development Workflows

Smart Contract Development:

cd contracts
forge build               # Compile contracts
forge test                # Run contract tests

Node Development:

cd operator
cargo build --release --features fast-runtime
cargo test
./scripts/run-benchmarks.sh

After Making Changes:

cd test
bun generate:wagmi        # Regenerate contract bindings
bun generate:types        # Regenerate runtime types

Key Features

Verifiable Decentralized Storage

Production-scale storage with cryptographic guarantees:

• Buckets: User-created containers managed by an MSP, summarized by a Merkle-Patricia trie root on-chain
• Files: Deterministically chunked, hashed into Merkle trees, with roots serving as immutable fingerprints
• Proofs: Merkle proofs enable verification of data integrity without trusting intermediaries
• Audits: BSPs prove ongoing data custody via randomized proof challenges

Storage Provider Network

Two-tier provider model balancing performance and reliability:

• MSPs: User-selected providers offering data retrieval with competitive service offerings
• BSPs: Network-assigned backup providers ensuring data redundancy and availability, with on-chain slashing for failed proof challenges
• Fisherman: Auditing service that monitors proofs and triggers challenges for misbehavior
• Indexer: Indexes on-chain storage events for efficient querying

EigenLayer Security

DataHaven validators secured through Ethereum restaking:

• Validators register as operators via DataHavenServiceManager contract
• Economic security through ETH restaking
• Slashing for validator misbehavior (separate from BSP slashing which is on-chain)
• Performance-based validator rewards through RewardsRegistry

EVM Compatibility

Full Ethereum Virtual Machine support via Frontier pallets:

• Deploy Solidity smart contracts
• Use existing Ethereum tooling (MetaMask, Hardhat, etc.)
• Compatible with ERC-20, ERC-721, and other standards

Cross-chain Communication

Trustless bridging via Snowbridge:

• Native token transfers between Ethereum ↔ DataHaven
• Cross-chain message passing
• Finality proofs via BEEFY consensus
• Three specialized relayers (beacon, BEEFY, execution)

Use Cases

DataHaven is designed for applications requiring verifiable, tamper-proof data storage:

• AI & Machine Learning: Store training datasets, model weights, and agent configurations with cryptographic proofs of integrity — enabling federated learning and verifiable AI pipelines
• DePIN (Decentralized Physical Infrastructure): Persistent storage for IoT sensor data, device configurations, and operational logs with provable data lineage
• Real World Assets (RWAs): Immutable storage for asset documentation, ownership records, and compliance data with on-chain verification

Docker Images

Production images published to DockerHub.

Build optimizations:

• sccache - Rust compilation caching
• cargo-chef - Dependency layer caching
• BuildKit cache mounts - External cache restoration

Build locally:

cd test
bun build:docker:operator    # Creates datahavenxyz/datahaven:local

Development Environment

VS Code Configuration

IDE configurations are excluded from version control for personalization, but these settings are recommended for optimal developer experience. Add to your .vscode/settings.json:

Rust Analyzer:

{
  "rust-analyzer.linkedProjects": ["./operator/Cargo.toml"],
  "rust-analyzer.cargo.allTargets": true,
  "rust-analyzer.procMacro.enable": false,
  "rust-analyzer.server.extraEnv": {
    "CARGO_TARGET_DIR": "target/.rust-analyzer",
    "SKIP_WASM_BUILD": 1
  },
  "rust-analyzer.diagnostics.disabled": ["unresolved-macro-call"],
  "rust-analyzer.cargo.buildScripts.enable": false
}

Optimizations:

• Links operator/ directory as the primary Rust project
• Disables proc macros and build scripts for faster analysis (Substrate macros are slow)
• Uses dedicated target directory to avoid conflicts
• Skips WASM builds during development

Solidity (Juan Blanco's extension):

{
  "solidity.formatter": "forge",
  "solidity.compileUsingRemoteVersion": "v0.8.28+commit.7893614a",
  "[solidity]": {
    "editor.defaultFormatter": "JuanBlanco.solidity"
  }
}

Note: Solidity version must match foundry.toml

TypeScript (Biome):

{
  "biome.lsp.bin": "test/node_modules/.bin/biome",
  "[typescript]": {
    "editor.defaultFormatter": "biomejs.biome",
    "editor.codeActionsOnSave": {
      "source.organizeImports.biome": "always"
    }
  }
}

CI/CD

Local CI Testing

Run GitHub Actions workflows locally using act:

# Run E2E workflow
act -W .github/workflows/e2e.yml -s GITHUB_TOKEN="$(gh auth token)"

# Run specific job
act -W .github/workflows/e2e.yml -j test-job-name

Automated Workflows

The repository includes GitHub Actions for:

• E2E Testing: Full integration tests on PR and main branch
• Contract Testing: Foundry test suites for smart contracts
• Rust Testing: Unit and integration tests for operator
• Docker Builds: Multi-platform image builds with caching
• Release Automation: Version tagging and changelog generation

See .github/workflows/ for workflow definitions.

Contributing

Development Cycle

1. Make Changes: Edit contracts, runtime, or tests
2. Run Tests: Component-specific tests (forge test, cargo test)
3. Regenerate Types: Update bindings if contracts/runtime changed
4. Integration Test: Run E2E tests to verify cross-component behavior
5. Code Quality: Format and lint (cargo fmt, forge fmt, bun fmt:fix)

Common Pitfalls

• Type mismatches: Regenerate with bun generate:types after runtime changes
• Contract changes not reflected: Run bun generate:wagmi after modifications
• Kurtosis issues: Ensure Docker is running and Kurtosis engine is started
• Slow development: Use --features fast-runtime for shorter epochs/eras (block time stays 6s)
• Network launch hangs: Check Blockscout - forge output can appear frozen

See CLAUDE.md for detailed development guidance.

License

GPL-3.0 - See LICENSE file for details

Links

• DataHaven Website
• DataHaven Documentation
• StorageHub Repository
• EigenLayer Documentation
• Substrate Documentation
• Snowbridge Documentation
• Foundry Book
• Polkadot-API Documentation