1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158

// Copyright (c) 2021, Facebook, Inc. and its affiliates
// Copyright (c) 2022, Mysten Labs, Inc.
// SPDX-License-Identifier: Apache-2.0
use config::{Committee, SharedWorkerCache, Stake, WorkerId};
use crypto::PublicKey;
use futures::stream::{futures_unordered::FuturesUnordered, StreamExt as _};
use network::{CancelOnDropHandler, P2pNetwork, ReliableNetwork};
use tokio::{sync::watch, task::JoinHandle};
use types::{
    error::DagError,
    metered_channel::{Receiver, Sender},
    Batch, ReconfigureNotification, WorkerMessage,
};

#[cfg(test)]
#[path = "tests/quorum_waiter_tests.rs"]
pub mod quorum_waiter_tests;

/// The QuorumWaiter waits for 2f authorities to acknowledge reception of a batch.
pub struct QuorumWaiter {
    /// The public key of this authority.
    name: PublicKey,
    /// The id of this worker.
    id: WorkerId,
    /// The committee information.
    committee: Committee,
    /// The worker information cache.
    worker_cache: SharedWorkerCache,
    /// Receive reconfiguration updates.
    rx_reconfigure: watch::Receiver<ReconfigureNotification>,
    /// Input Channel to receive commands.
    rx_message: Receiver<Batch>,
    /// Channel to deliver batches for which we have enough acknowledgments.
    tx_batch: Sender<Batch>,
    /// A network sender to broadcast the batches to the other workers.
    network: P2pNetwork,
}

impl QuorumWaiter {
    /// Spawn a new QuorumWaiter.
    #[must_use]
    pub fn spawn(
        name: PublicKey,
        id: WorkerId,
        committee: Committee,
        worker_cache: SharedWorkerCache,
        rx_reconfigure: watch::Receiver<ReconfigureNotification>,
        rx_message: Receiver<Batch>,
        tx_batch: Sender<Batch>,
        network: P2pNetwork,
    ) -> JoinHandle<()> {
        tokio::spawn(async move {
            Self {
                name,
                id,
                committee,
                worker_cache,
                rx_reconfigure,
                rx_message,
                tx_batch,
                network,
            }
            .run()
            .await;
        })
    }

    /// Helper function. It waits for a future to complete and then delivers a value.
    async fn waiter(
        wait_for: CancelOnDropHandler<anemo::Result<anemo::Response<()>>>,
        deliver: Stake,
    ) -> Stake {
        let _ = wait_for.await;
        deliver
    }

    /// Main loop.
    async fn run(&mut self) {
        loop {
            tokio::select! {
                Some(batch) = self.rx_message.recv() => {
                    // Broadcast the batch to the other workers.
                    let workers: Vec<_> = self
                        .worker_cache
                        .load()
                        .others_workers(&self.name, &self.id)
                        .into_iter()
                        .map(|(name, info)| (name, info.name))
                        .collect();
                    let (primary_names, worker_names): (Vec<_>, _) = workers.into_iter().unzip();
                    let message = WorkerMessage::Batch(batch.clone());
                    let handlers = self.network.broadcast(worker_names, &message).await;

                    // Collect all the handlers to receive acknowledgements.
                    let mut wait_for_quorum: FuturesUnordered<_> = primary_names
                        .into_iter()
                        .zip(handlers.into_iter())
                        .map(|(name, handler)| {
                            let stake = self.committee.stake(&name);
                            Self::waiter(handler, stake)
                        })
                        .collect();

                    // Wait for the first 2f nodes to send back an Ack. Then we consider the batch
                    // delivered and we send its digest to the primary (that will include it into
                    // the dag). This should reduce the amount of synching.
                    let threshold = self.committee.quorum_threshold();
                    let mut total_stake = self.committee.stake(&self.name);
                    loop {
                        tokio::select! {
                            Some(stake) = wait_for_quorum.next() => {
                                total_stake += stake;
                                if total_stake >= threshold {
                                    if self.tx_batch.send(batch).await.is_err() {
                                        tracing::debug!("{}", DagError::ShuttingDown);
                                    }
                                    break;
                                }
                            }

                            result = self.rx_reconfigure.changed() => {
                                result.expect("Committee channel dropped");
                                let message = self.rx_reconfigure.borrow().clone();
                                match message {
                                    ReconfigureNotification::NewEpoch(new_committee)
                                        | ReconfigureNotification::UpdateCommittee(new_committee) => {
                                            self.network.cleanup(self.committee.network_diff(&new_committee));
                                            self.committee = new_committee;
                                            tracing::debug!("Dropping batch: committee updated to {}", self.committee);
                                            break; // Don't wait for acknowledgements.
                                    },
                                    ReconfigureNotification::Shutdown => return
                                }
                            }
                        }
                    }
                },

                // Trigger reconfigure.
                result = self.rx_reconfigure.changed() => {
                    result.expect("Committee channel dropped");
                    let message = self.rx_reconfigure.borrow().clone();
                    match message {
                        ReconfigureNotification::NewEpoch(new_committee) => {
                            self.committee = new_committee;
                        },
                        ReconfigureNotification::UpdateCommittee(new_committee) => {
                            self.committee = new_committee;

                        },
                        ReconfigureNotification::Shutdown => return
                    }
                    tracing::debug!("Committee updated to {}", self.committee);
                }
            }
        }
    }
}