a16z/helios

Helios

Helios is a fully trustless, efficient, and portable Ethereum light client written in Rust.

Helios converts an untrusted centralized RPC endpoint into a safe unmanipulable local RPC for its users. It syncs in seconds, requires no storage, and is lightweight enough to run on mobile devices.

The entire size of Helios's binary is 5.3Mb and should be easy to compile into WebAssembly. This makes it a perfect target to embed directly inside wallets and dapps.

Installing

First install heliosup, Helios's installer:

curl https://raw.githubusercontent.com/a16z/helios/master/heliosup/install | bash

To install Helios, run heliosup.

Testing

To ensure that Helios works as expected, we have a comprehensive test suite that you can run. Before running the tests, make sure to create a .env file in the root of the project directory. You can copy the contents of the .env.example file and fill in your own secrets.

To run all tests, use the following command:

cargo test-all

To run tests for an individual package, use this command, replacing with the package you want to test:

cargo test -p <package-name>

Usage

To run Helios, run the below command, replacing $ETH_RPC_URL with an RPC provider URL such as Alchemy or Infura:

helios --execution-rpc $ETH_RPC_URL

$ETH_RPC_URL must be an Ethereum provider that supports the eth_getProof endpoint. Infura does not currently support this. We recommend using Alchemy.

Helios will now run a local RPC server at http://127.0.0.1:8545.

Helios provides examples in the examples/ directory. To run an example, you can execute cargo run -p helios --example <example_name> from inside the helios repository.

Helios also provides documentation of its supported RPC methods in the rpc.md file.

Warning

Helios is still experimental software. While we hope you try it out, we do not suggest adding it as your main RPC in wallets yet. Sending high-value transactions from a wallet connected to Helios is discouraged.

Additional Options

--consensus-rpc or -c can be used to set a custom consensus layer rpc endpoint. This must be a consensus node that supports the light client beaconchain api. We recommend using Nimbus for this. If no consensus rpc is supplied, it defaults to https://www.lightclientdata.org which is run by us.

--checkpoint or -w can be used to set a custom weak subjectivity checkpoint. This must be equal the first beacon blockhash of an epoch. Weak subjectivity checkpoints are the root of trust in the system. If this is set to a malicious value, an attacker can cause the client to sync to the wrong chain. Helios sets a default value initially, then caches the most recent finalized block it has seen for later use.

--network or -n sets the network to sync to. Current valid options are mainnet and goerli, however users can add custom networks in their configuration files.

--rpc-port or -p sets the port that the local RPC should run on. The default value is 8545.

--data-dir or -d sets the directory that Helios should use to store cached weak subjectivity checkpoints in. Each network only stores the latest checkpoint, which is just 32 bytes.

--fallback or -f sets the checkpoint fallback url (a string). This is only used if the checkpoint provided by the --checkpoint flag is too outdated for Helios to use to sync. If none is provided and the --load-external-fallback flag is not set, Helios will error. For example, you can specify the fallback like so: helios --fallback "https://sync-mainnet.beaconcha.in" (or using shorthand like so: helios -f "https://sync-mainnet.beaconcha.in")

--load-external-fallback or -l enables weak subjectivity checkpoint fallback (no value needed). For example, say you set a checkpoint value that is too outdated and Helios cannot sync to it. If this flag is set, Helios will query all network apis in the community-maintained list at ethpandaops/checkpoint-synz-health-checks for their latest slots. The list of slots is filtered for healthy apis and the most frequent checkpoint occurring in the latest epoch will be returned. Note: this is a community-maintained list and thus no security guarantees are provided. Use this is a last resort if your checkpoint passed into --checkpoint fails. This is not recommended as malicious checkpoints can be returned from the listed apis, even if they are considered healthy. This can be run like so: helios --load-external-fallback (or helios -l with the shorthand).

--strict-checkpoint-age or -s enables strict checkpoint age checking. If the checkpoint is over two weeks old and this flag is enabled, Helios will error. Without this flag, Helios will instead surface a warning to the user and continue. If the checkpoint is greater than two weeks old, there are theoretical attacks that can cause Helios and over light clients to sync incorrectly. These attacks are complex and expensive, so Helios disables this by default.

--help or -h prints the help message.

Configuration Files

All configuration options can be set on a per-network level in ~/.helios/helios.toml. Here is an example config file:

[mainnet]
consensus_rpc = "https://www.lightclientdata.org"
execution_rpc = "https://eth-mainnet.g.alchemy.com/v2/XXXXX"
checkpoint = "0x85e6151a246e8fdba36db27a0c7678a575346272fe978c9281e13a8b26cdfa68"

[goerli]
consensus_rpc = "http://testing.prater.beacon-api.nimbus.team"
execution_rpc = "https://eth-goerli.g.alchemy.com/v2/XXXXX"
checkpoint = "0xb5c375696913865d7c0e166d87bc7c772b6210dc9edf149f4c7ddc6da0dd4495"

A comprehensive breakdown of config options is available in the config.md file.

Using Helios as a Library

Helios can be imported into any Rust project. Helios requires the Rust nightly toolchain to compile.

use std::{str::FromStr, env};

use helios::{client::ClientBuilder, config::networks::Network, types::BlockTag};
use ethers::{types::Address, utils};
use eyre::Result;

#[tokio::main]
async fn main() -> Result<()> {
    let untrusted_rpc_url = env::var("UNTRUSTED_RPC_URL")?;

    let mut client = ClientBuilder::new()
        .network(Network::MAINNET)
        .consensus_rpc("https://www.lightclientdata.org")
        .execution_rpc(&untrusted_rpc_url)
        .build()?;

    client.start().await?;

    let head_block_num = client.get_block_number().await?;
    let addr = Address::from_str("0x00000000219ab540356cBB839Cbe05303d7705Fa")?;
    let block = BlockTag::Latest;
    let balance = client.get_balance(&addr, block).await?;

    println!("synced up to block: {}", head_block_num);
    println!("balance of deposit contract: {}", utils::format_ether(balance));

    Ok(())
}

Below we demonstrate fetching checkpoints from the community-maintained list of checkpoint sync apis maintained by ethPandaOps.

use eyre::Result;
use helios::config::{checkpoints, networks};

#[tokio::main]
async fn main() -> Result<()> {
    // Construct the checkpoint fallback services
    let cf = checkpoints::CheckpointFallback::new().build().await.unwrap();

    // Fetch the latest goerli checkpoint
    let goerli_checkpoint = cf.fetch_latest_checkpoint(&networks::Network::GOERLI).await.unwrap();
    println!("Fetched latest goerli checkpoint: {}", goerli_checkpoint);

    // Fetch the latest mainnet checkpoint
    let mainnet_checkpoint = cf.fetch_latest_checkpoint(&networks::Network::MAINNET).await.unwrap();
    println!("Fetched latest mainnet checkpoint: {}", mainnet_checkpoint);

    Ok(())
}

Architecture

graph LR

Client ----> Rpc
Client ----> Node
Node ----> ConsensusClient
Node ----> ExecutionClient
ExecutionClient ----> ExecutionRpc
ConsensusClient ----> ConsensusRpc
Node ----> Evm
Evm ----> ExecutionClient
ExecutionRpc --> UntrustedExecutionRpc
ConsensusRpc --> UntrustedConsensusRpc

classDef node fill:#f9f,stroke:#333,stroke-width:4px, color:black;
class Node,Client node
classDef execution fill:#f0f,stroke:#333,stroke-width:4px;
class ExecutionClient,ExecutionRpc execution
classDef consensus fill:#ff0,stroke:#333,stroke-width:4px;
class ConsensusClient,ConsensusRpc consensus
classDef evm fill:#0ff,stroke:#333,stroke-width:4px;
class Evm evm
classDef providerC fill:#ffc
class UntrustedConsensusRpc providerC
classDef providerE fill:#fbf
class UntrustedExecutionRpc providerE
classDef rpc fill:#e10
class Rpc rpc


subgraph "External Network"
UntrustedExecutionRpc
UntrustedConsensusRpc
end

Benchmarks

Benchmarks are defined in the benches subdirectory. They are built using the criterion statistics-driven benchmarking library.

To run all benchmarks, you can use cargo bench. To run a specific benchmark, you can use cargo bench --bench <name>, where <name> is one of the benchmarks defined in the Cargo.toml file under a [[bench]] section.

To learn more about helios benchmarking and to view benchmark flamegraphs, view the benchmark readme.

Contributing

All contributions to Helios are welcome. Before opening a PR, please submit an issue detailing the bug or feature. When opening a PR, please ensure that your contribution builds on the nightly rust toolchain, has been linted with cargo fmt, and contains tests when applicable.

Telegram

If you are having trouble with Helios or are considering contributing, feel free to join our telegram here.

Disclaimer

This code is being provided as is. No guarantee, representation or warranty is being made, express or implied, as to the safety or correctness of the code. It has not been audited and as such there can be no assurance it will work as intended, and users may experience delays, failures, errors, omissions or loss of transmitted information. Nothing in this repo should be construed as investment advice or legal advice for any particular facts or circumstances and is not meant to replace competent counsel. It is strongly advised for you to contact a reputable attorney in your jurisdiction for any questions or concerns with respect thereto. a16z is not liable for any use of the foregoing, and users should proceed with caution and use at their own risk. See a16z.com/disclosures for more info.