Skip to content

Latest commit

 

History

History
118 lines (90 loc) · 4.51 KB

README.md

File metadata and controls

118 lines (90 loc) · 4.51 KB
Raw

Staking Indexer

The staking indexer is a tool that extracts BTC staking relevant data from the Bitcoin blockchain, ensures that it follows the pre-requisites for a valid staking transaction, and determines whether the transaction should be active or not. All valid staking transactions are transformed into a structured form, stored in a database, and published as events in a RabbitMQ messaging queue for consumption by consumers. The staking indexer is the enforcer of the Bitcoin Staking protocol and serves as the ground truth for the Bitcoin Staking system.

Features

  1. Polling BTC blocks data from a specified height in an ongoing manner. The poller ensures that all the output blocks have at least N confirmations where N is a configurable value, which should be large enough so that the chance of the output blocks being forked is enormously low, e.g., greater than or equal to 6 in Bitcoin mainnet. In case of major reorg, the indexer will terminate and should manually bootstrap from a clean DB.
  2. Extracting transaction data for staking, unbonding, and withdrawal. These transactions are verified and compared against the system parameters to identify whether they are active, inactive due to staking cap overflow, or invalid. The details of the protocol for verifying and activating transactions can be found here.
  3. Calculating confirmed and unconfirmed TVL (total value locked) based on observed transactions.
  4. Storing the extracted transaction data and system state in a database. The details can be found here.
  5. Pushing staking, unbonding, withdrawal events, and TVL calculation results to the message queues. A reference implementation based on rabbitmq is provided. The definition of each type of events can be found here. Our API service exhibits how these events are utilized and presented.
  6. Monitoring the status of the service through Prometheus metrics.

Usage

1. Setup bitcoind node

The staking indexer relies on bitcoind as backend. Follow this guide to set up a bitcoind node.

2. Install

Clone the repository to your local machine from Github:

git clone https://github.com/babylonchain/staking-indexer.git

Install the sid daemon binary by running:

cd staking-indexer # cd into the project directory
make install

3. Configuration

To initiate the program with default config file, run:

sid init

This will create a sid.conf file in the default home directory. The default home directories for different operating systems are:

  • MacOS ~/Users/<username>/Library/Application Support/Sid
  • Linux ~/.Sid
  • Windows C:\Users\<username>\AppData\Local\Sid

Use the --home flag to specify the home directory and use the --force to overwrite the existing config file.

4. Run the Staking Indexer

To run the staking indexer, we need to prepare a global-params.json file which defines all the global params that are used across the BTC staking system. The indexer needs it to parse staking transaction data. The definition of global params can be found here. An example of the global params can be found in test-params.json. The program reads the file from the home directory by default. The user can specify the file path using the --params-path flag.

To run the staking indexer from a specific height, run:

sid start --start-height <start-height>

If the --start-height is not specified, the indexer will retrieve the start height first from the database which saves the last_processed_height. If the database is empty, the start height will be retrieved from the earliest activation_height defined in the global parameters file. The earliest activation_height is a height before which no staking transactions have been included in. Note that if the database is empty, the indexer will strictly start from the earliest activation_height. If the database is not empty, the user can specify a height that is not higher than last_processed_height + 1 via --start-height. This is to ensure that no staking data will be missed.

Tests

Run unit tests:

make test

Run e2e tests:

make test-e2e

This will initiate docker containers for both a bitcoind node running in the signet mode and a rabbitmq instance.