The Espresso Sequencer offers rollups credible neutrality and enhanced interoperability, without compromising on scale. Consisting of a data availability solution and a decentralized network of nodes that sequences transactions, layer-2 rollups can leverage the Espresso Sequencer to give developers and end users fast confirmations, low (and fair) fees, and robust infrastructure.
This diagram below depicts a simplified view of the current architecture of the Espresso Sequencer. The diagram includes views of an Espresso Sequencer node, the Espresso Sequencer Network (nodes, CDN, builders, prover, state relay service), two rollups (one ZK rollup "Z", one optimistic rollup "O") that use the Espresso Sequencer for sequencing and some important L1 contracts.
- Glossary
- Namespace: an identifier to distinguish rollups, akin to an Ethereum chain ID
- Rollup transaction: a transaction a user submits to a rollup, usually an EVM transaction
- Transaction: a transaction inside the Espresso Sequencer: a rollup transaction plus a namespace ID of the rollup
- Rollup block: a block in a rollup consisting only of transactions in this rollup
- Espresso block: a block produced by the Espresso sequencer containing transactions of multiple rollups
The sequence diagram below serves as a complement to the architecture diagram. The following interactions are depicted.
- Builders deposit funds into the fee contract on Ethereum Layer 1. These funds are later used to pay fees.
- Users submit transactions to the Submit APIs of sequencer nodes.
- The leader/proposer obtains a block from a builder.
- HotShot consensus creates new blocks containing sequenced rollup transactions.
- A rollup produces a rollup block with transactions sequenced by the Espresso sequencer.
- A proof for a HotShot state update is created and verified in the Light Client smart contract.
- A ZK rollup proves a correct state transaction by sending a proof to its rollup smart contract.
- A dispute is settled in an optimistic rollup. If necessary, the HotShot commitment is read from the Light Client contract.
In order for ZK rollups to rely on blocks produced by Espresso as a source of transactions, it is required to adjust the circuit that encodes the state update logic. See zk-rollups integration for more details.
Refer to sequencer-example-l2 for instructions on how to run a dockerized Espresso Sequencer network with an example Layer 2 rollup application.
- Obtain code:
git clone [email protected]:EspressoSystems/espresso-sequencer
. - Make sure nix is installed.
- Activate the environment with
nix-shell
, ornix develop
, ordirenv allow
if using direnv. - For installation without nix please see ubuntu.md.
The rust code documentation can be found at sequencer.docs.espressosys.com. Please note the disclaimer about API stability at the end of the readme.
To generate the documentation locally and view it in the browser, run
just doc --open
just pull # to pull docker images
just test
make doc
Docker images and the docker-compose-demo.yaml file are provided for convenience. The
Docker-based demo fetches the images from the ghcr
repository, where they are updated with every push to main
on
GitHub. For testing uncommitted changes, you can also run the binaries by manually building and running the services.
Build all executables with cargo build --release
. You may then start a sequencer network. First, start an
orchestrator. Choose a port $PORT
to run it on and decide how many sequencer nodes $N
you will use, then run
target/release/orchestrator -p $PORT -n $N
.
The sequencer will distribute a HotShot configuration to all the nodes which connect to it, which specifies consensus
parameters like view timers. There is a default config, but you can override any parameters you want by passing
additional options to the orchestrator
executable. Run target/release/orchestrator --help
to see a list of available
options.
Next, you must launch a cdn
instance, which is necessary to facilitate consensus.
just dev-cdn -- -p 1738
In this case, we run it on port 1738.
Once you have started the orchestrator and the CDN, you must connect $N
sequencer nodes to them, after which the
network will start up automatically. To start one node, run
target/release/sequencer \
--orchestrator-url http://localhost:$PORT \
--cdn-endpoint "127.0.0.1:1738" \
-- http --port 8083 -- query --storage-path storage -- submit
A useful Bash snippet for running $N
nodes simultaneously in the background of your shell is:
for i in `seq $N`; do
target/release/sequencer \
--orchestrator-url http://localhost:$PORT \
--cdn-endpoint "127.0.0.1:1738" \
done
For running a full demo natively run just demo-native
.
A foundry project for the contracts specific to HotShot can be found in the directory contracts
.
To compile
forge build
To run the tests
forge test
In order to avoid constant warnings about checksum mismatches with svm-rs
managed solc
we set FOUNDRY_SRC
to solc installed via flake.nix.
- To use the contracts from rust generate the rust contracts bindings:
just gen-bindings
. - Bindings are only generated for contracts in the
contracts/src
folder
To generate documentation in ./docs
for solidity code run
forge doc
To deploy the contracts to a local testnet, first run a dev chain (e.g. anvil
), then run
forge script DeployHotShot --broadcast --rpc-url local
To deploy to sepolia set SEPOLIA_RPC_URL
and MNEMONIC
env vars and run
forge script DeployHotShot --broadcast --rpc-url sepolia
To additionally verify the contact on etherscan set the ETHERSCAN_API_KEY
env var and run
forge script DeployHotShot --broadcast --rpc-url sepolia --verify
Running the script will save a file with details about the deployment in contracts/broadcast/$CHAIN_ID
.
- code for demo purposes goes into the
contracts/demo
folder - code that eventually ends up in production goes into the
contracts/src
folder
The gas consumption for verifying a plonk proof as well as updating the state of the light client contract can be seen by running:
> just gas-benchmarks
> cat gas-benchmarks.txt
[PASS] test_verify_succeeds() (gas: 507774)
[PASS] testCorrectUpdateBench() (gas: 594533)
In order to profile the gas consumption of the light client contract do the following:
- Set the environment variables
SEPOLIA_RPC_URL
,MNEMONIC
andETHERSCAN_API_KEY
. just lc-contract-profiling-sepolia
- Create an account on sentio.xyz.
- Use the hash of the transaction generated in step two when calling the function
newFinalizedState
in order to obtain the gas profile.
This is only necessary to fetch private images.
- Go to your github profile
- Developer Settings > Personal access tokens > Personal access tokens (classic)
- Generate a new token
- for the scope options of the token, tick the repo box.
- Run
docker login ghcr.io --username <you_github_id> --password <your_personal_access_token>
(c) 2022 Espresso Systems espresso-sequencer
was developed by Espresso Systems. While we plan to adopt an open
source license, we have not yet selected one. As such, all rights are reserved for the time being. Please reach out to
us if you have thoughts on licensing.
DISCLAIMER: This software is provided "as is" and its security has not been externally audited. Use at your own risk.
DISCLAIMER: The Rust library crates provided in this repository are intended primarily for use by the binary targets in this repository. We make no guarantees of public API stability. If you are building on these crates, reach out by opening an issue to discuss the APIs you need.