This document explains everything a validator operator should know to run a validator. It intentionally skips advanced features on how users can use the protocol or any user-experience-related information.

What is a validator?

Validators are the execution unit/actors of the protocol.

They have the following responsibilities:

• Listen to on-chain events to materialize Tableland-compliant SQL queries in a database engine (currently, SQLite by default).
• Serve read-queries (e.g: SELECT * FROM foo_69_1) to the external world.

<aside> 💡 The responsibilities of the validator will continue to change as the Tableland protocol evolves. In the future, validators will have more responsibilities in the network.

</aside>

Where does the validator fit in the Tableland protocol?

The following is a diagram that describes at a high level the interaction between the validator, EVM-chains, and the external world:

(Excalidraw link)

To understand better the usual work mechanics of the validator, let’s go through a typical use-case where a user mints a table, adds data to the table and reads it:

1- The user will mint a table in the Registry smart contract.

2- The Registry contract will emit a CreateTable event containing the CREATE TABLE statement as extra data.

3- Validators will detect the new event and execute the CREATE TABLE statement.