Silk, a silky smooth implementation of the Loom specification

Loom is a new achitecture for a high performance blockchain. Its whitepaper boasts a theoretical throughput of 710k transactions per second on a 1 gbps network. The specification is implemented in two git repositories. Reserach is performed in the loom repository. That work drives the Loom specification forward. This repository, on the other hand, aims to implement the specification as-is. We care a great deal about quality, clarity and short learning curve. We avoid the use of unsafe Rust and write tests for everything. Optimizations are only added when corresponding benchmarks are also added that demonstrate real performance boosts. We expect the feature set here will always be a ways behind the loom repo, but that this is an implementation you can take to the bank, literally.

Usage

Add the latest silk package to the [dependencies] section of your Cargo.toml.

Create a Historian and send it events to generate an event log, where each log entry is tagged with the historian's latest hash. Then ensure the order of events was not tampered with by verifying each entry's hash can be generated from the hash in the previous entry:

extern crate silk;

use silk::historian::Historian;
use silk::log::{verify_slice, Entry, Event, Sha256Hash};
use std::thread::sleep;
use std::time::Duration;
use std::sync::mpsc::SendError;

fn create_log(hist: &Historian) -> Result<(), SendError<Event>> {
    sleep(Duration::from_millis(15));
    hist.sender.send(Event::UserDataKey(Sha256Hash::default()))?;
    sleep(Duration::from_millis(10));
    Ok(())
}

fn main() {
    let seed = Sha256Hash::default();
    let hist = Historian::new(&seed, Some(10));
    create_log(&hist).expect("send error");
    drop(hist.sender);
    let entries: Vec<Entry> = hist.receiver.iter().collect();
    for entry in &entries {
        println!("{:?}", entry);
    }

    // Proof-of-History: Verify the historian learned about the events
    // in the same order they appear in the vector.
    assert!(verify_slice(&entries, &seed));
}

Running the program should produce a log similar to:

Entry { num_hashes: 0, end_hash: [0, ...], event: Tick }
Entry { num_hashes: 2, end_hash: [67, ...], event: UserDataKey(3735928559) }
Entry { num_hashes: 3, end_hash: [123, ...], event: Tick }

Proof-of-History

Take note of the last line:

assert!(verify_slice(&entries, &seed));

It's a proof! For each entry returned by the historian, we can verify that end_hash is the result of applying a sha256 hash to the previous end_hash exactly num_hashes times, and then hashing then event data on top of that. Because the event data is included in the hash, the events cannot be reordered without regenerating all the hashes.

Developing

Building

Install rustc, cargo and rustfmt:

$ curl https://sh.rustup.rs -sSf | sh
$ source $HOME/.cargo/env
$ rustup component add rustfmt-preview

Download the source code:

$ git clone https://github.com/loomprotocol/silk.git
$ cd silk

Testing

Run the test suite:

cargo test

Benchmarking

First install the nightly build of rustc. cargo bench requires unstable features:

$ rustup install nightly

Run the benchmarks:

$ cargo +nightly bench --features="asm,unstable"