Overview

The rollup node is responsible for maintaining the L2 chain from transactions included in the L1 chain and directly sequenced.

The rollup node operates in several modes:

• Verifier: Constructs L2 directly from L1. Will be behind the L2 head.
• Sequencer: Constructs L2 blocks from incoming transactions & deposits. Runs ahead of what is submitted on chain.
• Batch Submitter: Posts L2 blocks to L1 before the Sequence Window is up.
• Output Submitter: Posts the result of L2 execution to L1 on a regular schedule.

Components

• eth: Definitions of common structures
• l1: Wraps the Ethereum Client API to provide a simple download API as well as expose certain client methods
• l2: Wraps the Ethereum Client Engine API to provide payload execution and fork choise options.
• node: Runs the rollup node
• test: End to end test of the rollup node
• rollup: Implements the logic of maintaining L2 from L1

Rollup Package Overview

The rollup driver maintains a view of the L1 and L2 chains. When an updates happens to either chain, it responds and updates the L2 chain appropriately.

Inside the driver, the state object maintains a view of the chains. In the case of a linear extension to L1, it update's its state and then if is able to, creates a list of what L1 blocks should be looked when creating the next l2 blocks and passes it to the step function. In the case of a re-org, it relies on the sync package to determine what the new L2 head should be, and then continues the process of syncing from the L2 Head & it's associated L1 Parent.

The step function takes the sequence window and the L2 head (what will become the L2 parent of the first L2 block) and then does the following:

• Fetch the L1 blocks in the specified range.
• Parse the L1 blocks to get the deposits and batches.
• Create the L2 block templates (multiple payload attributes) from the deposits and batches.
• Execute the L2 payload attributes to derive the finalized blocks
• Update the L2 Head (via the ForkChoice API)