Author: 1912212.eth, Foresight News

With the increasing maturity of Ethereum's second-layer network and the significant reduction of Gas fees, these networks are showing a trend of vigorous growth, but problems have also emerged, such as fees and speeds will be affected in many ways But it is not satisfactory, making the user experience unsatisfactory. In this context, the importance of transaction sequencing has become increasingly prominent, becoming the key to solving transaction bottlenecks and optimizing user experience.

In April this year, Astria, a modular blockchain focusing on shared sorters, completed a $5.5 million seed round of financing, led by Maven 11, with participation from 1kx, Delphi Ventures, Robot Ventures, etc. Just three months later, Astria completed another $12.5 million in financing, led by dba and Placeholder VC, with participation from Hasu and others.

What is Astria?

Astria is developing a decentralized shared sorter network designed to provide fast finality, censorship resistance, composability, and decentralization for Rollups.

Currently, it is more convenient, cheaper, and user-friendly for L2 to run a centralized sorter, so the mainstream L2 is run by its own team. Although L2 users can submit transactions directly to L1 to bypass the sorter, users must pay transaction gas fees to L1, and transactions may take longer to be completed.

The sorter controls the ordering of transactions and theoretically has the power to exclude user transactions from being included. The sorter can also extract MEV from the transaction group. If there is only one sorter, the risk of centralization will increase.

Therefore, a decentralized shared sorter still makes sense.

How Astria works

Astria's decentralized sorter has multiple sorter nodes, which can sort Rollup transactions. In Astria's operating mode, users submit transactions to Rollups, and the transactions automatically enter the memory pool of their respective Rollup nodes, where the combiner is responsible for collecting txnx and sending it to the sorter. Finally, the sorter aggregates the txnx into a shared block and finally sends a pre-confirmation to the user.

Current sorters are implemented based on specific Rollups. Astria batches blocks for multiple Rollups. Through data compression, more costs can be saved when publishing data to L1. The decentralized shared sorter network will incentivize participants from multiple Rollup ecosystems to act as validators on the network.

Astria Stack

The main components of Astria include 5 parts, namely, combiner, sorting layer, repeater, DA and scheduler.

Combiners

Professionals who understand technology may be able to directly use the sorting layer to perform better transaction sorting, but it will be more difficult for most ordinary users to actually use it. Directly interacting with the sorting layer requires users to hold sorter tokens and maintain sorter wallets, both of which have a significant negative impact on the user experience.

Astria provides users with tools to abstract this complexity, called combiners. A combiner is like a gas station that takes on the cost of sorting user transactions. The combiner also provides users with out-of-order guarantees, bundling transactions in the order they were received.

Sorting Layer

Astria sorting layer uses CometBFT as its consensus algorithm. A chain that supports CometBFT is able to support IBC (Inter-Blockchain Communication), which means it can cross-chain between many other chains.

The Astria sorter is unique in that the transactions it contains are not executed (delayed sorting) but are instead assigned to another execution engine, the Rollup. A sorter node can choose to act as a "validator", meaning it actively participates in the production and finalization of new blocks.

The Astria sequencer's application logic allows for three main functions:

  • Sorting of Rollup Data

  • Value Transfer

  • Changes to the validator set

Repeater

The function of the relayer is to take the validated blocks from the sorter and pass them to the DA layer. Since the sorter has a faster block time than the DA, the relayer batches the ordered data from multiple sorter blocks before compressing them and submitting them to the DA.

Individual sorter blocks can also be fetched by the scheduler before the relayer commits them to the DA. This enables fast finalization with an improved user experience, acting as a soft commit for the execution layer. The collection of data sent to the DA layer by the relayer is used as the source of truth and is eventually extracted from the DA to be used as a confirmed commit for finality in the Rollup.

AND

Astria uses Celestia as a data availability layer, which is the final destination for all data sorted by the sorter network. Once data is written to Celestia, the transaction order is considered final, and when a new Rollup node is started, all data will be pulled from here.

scheduler

The scheduler can be thought of as the consensus implementation of the Rollup full node, similar to the operation node in the OP Stack. The scheduler is the counterpart of the execution engine, and together they form a complete Rollup node. Its role is to connect the sorter and DA layer to the Rollup execution layer by extracting transactions belonging to the Rollup node from each sorter block and forwarding them to the execution layer.

For each sorter block, it extracts the relevant Rollup data it needs, then verifies the batch of Rollup data, and waits for the verification to be completed, it converts it into a list of transactions and passes it to the execution engine.