diff options
Diffstat (limited to 'src/peering/basalt.rs')
| -rw-r--r-- | src/peering/basalt.rs | 475 |
1 files changed, 475 insertions, 0 deletions
diff --git a/src/peering/basalt.rs b/src/peering/basalt.rs new file mode 100644 index 0000000..be807a8 --- /dev/null +++ b/src/peering/basalt.rs @@ -0,0 +1,475 @@ +use std::collections::HashSet; +use std::net::SocketAddr; +use std::sync::{Arc, RwLock}; +use std::time::Duration; + +use log::{debug, warn}; +use lru::LruCache; +use rand::{thread_rng, Rng}; +use serde::{Deserialize, Serialize}; + +use sodiumoxide::crypto::hash; +use sodiumoxide::crypto::sign::ed25519; + +use crate::conn::*; +use crate::message::*; +use crate::netapp::*; +use crate::proto::*; + +// -- Protocol messages -- + +#[derive(Serialize, Deserialize)] +struct PullMessage {} + +impl Message for PullMessage { + const KIND: MessageKind = 0x42001100; + type Response = PushMessage; +} + +#[derive(Serialize, Deserialize)] +struct PushMessage { + peers: Vec<Peer>, +} + +impl Message for PushMessage { + const KIND: MessageKind = 0x42001101; + type Response = (); +} + +// -- Algorithm data structures -- + +type Seed = [u8; 32]; + +#[derive(Hash, Clone, Copy, Debug, PartialOrd, PartialEq, Eq, Serialize, Deserialize)] +struct Peer { + id: ed25519::PublicKey, + addr: SocketAddr, +} + +type Cost = [u8; 40]; +const MAX_COST: Cost = [0xffu8; 40]; + +impl Peer { + fn cost(&self, seed: &Seed) -> Cost { + let mut hasher = hash::State::new(); + hasher.update(&seed[..]); + + let mut cost = [0u8; 40]; + match self.addr { + SocketAddr::V4(v4addr) => { + let v4ip = v4addr.ip().octets(); + + for i in 0..4 { + let mut h = hasher.clone(); + h.update(&v4ip[..i + 1]); + cost[i * 8..(i + 1) * 8].copy_from_slice(&h.finalize()[..8]); + } + } + SocketAddr::V6(v6addr) => { + let v6ip = v6addr.ip().octets(); + + for i in 0..4 { + let mut h = hasher.clone(); + h.update(&v6ip[..i + 2]); + cost[i * 8..(i + 1) * 8].copy_from_slice(&h.finalize()[..8]); + } + } + } + + { + let mut h5 = hasher.clone(); + h5.update(&format!("{}", self.addr).into_bytes()[..]); + cost[32..40].copy_from_slice(&h5.finalize()[..8]); + } + + cost + } +} + +struct BasaltSlot { + seed: Seed, + peer: Option<Peer>, +} + +impl BasaltSlot { + fn cost(&self) -> Cost { + self.peer.map(|p| p.cost(&self.seed)).unwrap_or(MAX_COST) + } +} + +struct BasaltView { + i_reset: usize, + slots: Vec<BasaltSlot>, +} + +impl BasaltView { + fn new(size: usize) -> Self { + let slots = (0..size) + .map(|_| BasaltSlot { + seed: rand_seed(), + peer: None, + }) + .collect::<Vec<_>>(); + Self { i_reset: 0, slots } + } + + fn current_peers(&self) -> HashSet<Peer> { + self.slots + .iter() + .filter(|s| s.peer.is_some()) + .map(|s| s.peer.unwrap().clone()) + .collect::<HashSet<_>>() + } + fn current_peers_vec(&self) -> Vec<Peer> { + self.current_peers().drain().collect::<Vec<_>>() + } + + fn sample(&self, count: usize) -> Vec<Peer> { + let possibles = self + .slots + .iter() + .enumerate() + .filter(|(_i, s)| s.peer.is_some()) + .map(|(i, _s)| i) + .collect::<Vec<_>>(); + if possibles.len() == 0 { + vec![] + } else { + let mut ret = vec![]; + let mut rng = thread_rng(); + for _i in 0..count { + let idx = rng.gen_range(0, possibles.len()); + ret.push(self.slots[possibles[idx]].peer.unwrap()); + } + ret + } + } + + fn update_slot(&mut self, i: usize, peers: &[Peer]) { + let mut slot_cost = self.slots[i].cost(); + + for peer in peers.iter() { + let peer_cost = peer.cost(&self.slots[i].seed); + if self.slots[i].peer.is_none() || peer_cost < slot_cost { + self.slots[i].peer = Some(*peer); + slot_cost = peer_cost; + } + } + } + fn update_all_slots(&mut self, peers: &[Peer]) { + for i in 0..self.slots.len() { + self.update_slot(i, peers); + } + } + + fn disconnected(&mut self, id: ed25519::PublicKey) { + let mut cleared_slots = vec![]; + for i in 0..self.slots.len() { + if let Some(p) = self.slots[i].peer { + if p.id == id { + self.slots[i].peer = None; + cleared_slots.push(i); + } + } + } + + let remaining_peers = self.current_peers_vec(); + + for i in cleared_slots { + self.update_slot(i, &remaining_peers[..]); + } + } + + fn should_try_list(&self, peers: &[Peer]) -> Vec<Peer> { + // Select peers that have lower cost than any of our slots + let mut ret = HashSet::new(); + + for i in 0..self.slots.len() { + if self.slots[i].peer.is_none() { + return peers.to_vec(); + } + let mut min_cost = self.slots[i].cost(); + let mut min_peer = None; + for peer in peers.iter() { + if ret.contains(peer) { + continue; + } + let peer_cost = peer.cost(&self.slots[i].seed); + if peer_cost < min_cost { + min_cost = peer_cost; + min_peer = Some(*peer); + } + } + if let Some(p) = min_peer { + ret.insert(p); + if ret.len() == peers.len() { + break; + } + } + } + + ret.drain().collect::<Vec<_>>() + } + + fn reset_some_slots(&mut self, count: usize) { + for _i in 0..count { + self.slots[self.i_reset].seed = rand_seed(); + self.i_reset = (self.i_reset + 1) % self.slots.len(); + } + } +} + +pub struct BasaltParams { + pub view_size: usize, + pub cache_size: usize, + pub exchange_interval: Duration, + pub reset_interval: Duration, + pub reset_count: usize, +} + +pub struct Basalt { + netapp: Arc<NetApp>, + + param: BasaltParams, + bootstrap_peers: Vec<Peer>, + + view: RwLock<BasaltView>, + current_attempts: RwLock<HashSet<Peer>>, + backlog: RwLock<LruCache<Peer, ()>>, +} + +impl Basalt { + pub fn new( + netapp: Arc<NetApp>, + bootstrap_list: Vec<(ed25519::PublicKey, SocketAddr)>, + param: BasaltParams, + ) -> Arc<Self> { + let bootstrap_peers = bootstrap_list + .iter() + .map(|(id, addr)| Peer { + id: *id, + addr: *addr, + }) + .collect::<Vec<_>>(); + + let view = BasaltView::new(param.view_size); + let backlog = LruCache::new(param.cache_size); + + let basalt = Arc::new(Self { + netapp: netapp.clone(), + param, + bootstrap_peers, + view: RwLock::new(view), + current_attempts: RwLock::new(HashSet::new()), + backlog: RwLock::new(backlog), + }); + + let basalt2 = basalt.clone(); + netapp.on_connected.store(Some(Arc::new(Box::new( + move |pk: ed25519::PublicKey, addr: SocketAddr, is_incoming: bool| { + basalt2.on_connected(pk, addr, is_incoming); + }, + )))); + + let basalt2 = basalt.clone(); + netapp.on_disconnected.store(Some(Arc::new(Box::new( + move |pk: ed25519::PublicKey, is_incoming: bool| { + basalt2.on_disconnected(pk, is_incoming); + }, + )))); + + let basalt2 = basalt.clone(); + netapp.add_msg_handler::<PullMessage, _, _>( + move |_from: ed25519::PublicKey, _pullmsg: PullMessage| { + let push_msg = basalt2.make_push_message(); + async move { Ok(push_msg) } + }, + ); + + let basalt2 = basalt.clone(); + netapp.add_msg_handler::<PushMessage, _, _>( + move |_from: ed25519::PublicKey, push_msg: PushMessage| { + basalt2.handle_peer_list(&push_msg.peers[..]); + async move { Ok(()) } + }, + ); + + basalt + } + + pub fn sample(&self, count: usize) -> Vec<ed25519::PublicKey> { + self.view + .read() + .unwrap() + .sample(count) + .iter() + .map(|p| p.id) + .collect::<Vec<_>>() + } + + pub async fn run(self: Arc<Self>) { + for peer in self.bootstrap_peers.iter() { + tokio::spawn(self.clone().try_connect(*peer)); + } + + let pushpull_loop = self.clone().run_pushpull_loop(); + let reset_loop = self.run_reset_loop(); + tokio::join!(pushpull_loop, reset_loop); + } + + async fn run_pushpull_loop(self: Arc<Self>) { + loop { + tokio::time::delay_for(self.param.exchange_interval).await; + + let peers = self.view.read().unwrap().sample(2); + if peers.len() == 2 { + let (c1, c2) = { + let client_conns = self.netapp.client_conns.read().unwrap(); + ( + client_conns.get(&peers[0].id).cloned(), + client_conns.get(&peers[1].id).cloned(), + ) + }; + if let Some(c) = c1 { + tokio::spawn(self.clone().do_pull(c)); + } + if let Some(c) = c2 { + tokio::spawn(self.clone().do_push(c)); + } + } + } + } + + async fn do_pull(self: Arc<Self>, peer: Arc<ClientConn>) { + match peer.request(PullMessage {}, prio::NORMAL).await { + Ok(resp) => { + self.handle_peer_list(&resp.peers[..]); + } + Err(e) => { + warn!("Error during pull exchange: {}", e); + } + }; + } + + async fn do_push(self: Arc<Self>, peer: Arc<ClientConn>) { + let push_msg = self.make_push_message(); + if let Err(e) = peer.request(push_msg, prio::NORMAL).await { + warn!("Error during push exchange: {}", e); + } + } + + fn make_push_message(&self) -> PushMessage { + let current_peers = self.view.read().unwrap().current_peers_vec(); + PushMessage { + peers: current_peers, + } + } + + async fn run_reset_loop(self: Arc<Self>) { + loop { + tokio::time::delay_for(self.param.reset_interval).await; + + { + let mut view = self.view.write().unwrap(); + let prev_peers = view.current_peers(); + let prev_peers_vec = prev_peers.iter().cloned().collect::<Vec<_>>(); + + view.reset_some_slots(self.param.reset_count); + view.update_all_slots(&prev_peers_vec[..]); + + let new_peers = view.current_peers(); + drop(view); + + self.close_all_diff(&prev_peers, &new_peers); + } + + let mut to_retry_maybe = self.bootstrap_peers.clone(); + for (peer, _) in self.backlog.read().unwrap().iter() { + if !self.bootstrap_peers.contains(peer) { + to_retry_maybe.push(*peer); + } + } + self.handle_peer_list(&to_retry_maybe[..]); + } + } + + fn handle_peer_list(self: &Arc<Self>, peers: &[Peer]) { + let to_connect = self.view.read().unwrap().should_try_list(peers); + + for peer in to_connect.iter() { + tokio::spawn(self.clone().try_connect(*peer)); + } + } + + async fn try_connect(self: Arc<Self>, peer: Peer) { + { + let view = self.view.read().unwrap(); + let mut attempts = self.current_attempts.write().unwrap(); + + if view.slots.iter().any(|x| x.peer == Some(peer)) { + return; + } + if attempts.contains(&peer) { + return; + } + + attempts.insert(peer); + } + let res = self.netapp.clone().try_connect(peer.addr, peer.id).await; + debug!("Connection attempt to {}: {:?}", peer.addr, res); + + self.current_attempts.write().unwrap().remove(&peer); + + if res.is_err() { + self.backlog.write().unwrap().pop(&peer); + } + } + + fn on_connected(self: &Arc<Self>, pk: ed25519::PublicKey, addr: SocketAddr, is_incoming: bool) { + if is_incoming { + self.handle_peer_list(&[Peer{id: pk, addr}][..]); + } else { + let peer = Peer { id: pk, addr }; + + let mut backlog = self.backlog.write().unwrap(); + if backlog.get(&peer).is_none() { + backlog.put(peer, ()); + } + drop(backlog); + + let mut view = self.view.write().unwrap(); + let prev_peers = view.current_peers(); + + view.update_all_slots(&[peer][..]); + + let new_peers = view.current_peers(); + drop(view); + + self.close_all_diff(&prev_peers, &new_peers); + } + } + + fn on_disconnected(&self, pk: ed25519::PublicKey, is_incoming: bool) { + if !is_incoming { + self.view.write().unwrap().disconnected(pk); + } + } + + fn close_all_diff(&self, prev_peers: &HashSet<Peer>, new_peers: &HashSet<Peer>) { + let client_conns = self.netapp.client_conns.read().unwrap(); + for peer in prev_peers.iter() { + if !new_peers.contains(peer) { + if let Some(c) = client_conns.get(&peer.id) { + debug!("Closing connection to {} ({})", hex::encode(peer.id), peer.addr); + c.close(); + } + } + } + } +} + +fn rand_seed() -> Seed { + let mut seed = [0u8; 32]; + sodiumoxide::randombytes::randombytes_into(&mut seed[..]); + seed +} |