Everything you need to start running a validator
Hyperlane validators are stateless, do not submit transactions, and are not networked with other validators. Hyperlane validators are run on a per-origin-chain basis, and these instructions are written for a single chain.
Running a validator simply requires the following:
Validators make simple view calls to read merkle roots from the
Mailboxcontract on the chain they are validating for.
Operating a validator for Polygon mainnet requires access to an archive node. This is because validators should only sign roots once they've been finalized, and Polygon requires 256 block confirmations to achieve finality.
Validators use this key to sign the
Mailbox'slatest merkle root. Securing this key is important. If it is compromised, attackers can attempt to falsify messages, causing the validator to be slashed.
The Hyperlane validator agent currently supports signing with AWS KMS keys that are accessed via API keys/secrets as well as hexadecimal plaintext keys for testing. See more under Agent keys.
Validators write their signatures off-chain to publicly accessible, highly available, storage, so that they can be aggregated by relayers.
The Hyperlane validator agent currently supports storing signatures on AWS S3 using the same AWS API key above, as well as storing signatures in the local filesystem for testing.
Validators can compile the Rust binary themselves, or run a Docker image provided by Abacus Works. The binary is completely stateless and can be run using your favorite cloud service. You can even run multiple instances of them in different regions for high availability, as Hyperlane has no notion of "double signing".