project-serum
/
solana
mirror of https://github.com/project-serum/solana.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Add fork selection RFC (#2061) 

RFC and simulation for fork generation.
This commit is contained in:
anatoly yakovenko 2018-12-14 11:15:23 -08:00 committed by GitHub
parent 79f12d6b55
commit a22e1199cf
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 754 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
book/src/SUMMARY.md
View File

@ -26,6 +26,7 @@
- [Ledger Replication](ledger-replication.md)
- [Secure Enclave](enclave.md)
- [Staking Rewards](staking-rewards.md)
- [Fork Selection](fork-selection.md)
- [Entry Tree](entry-tree.md)
## Appendix

155
book/src/fork-selection.md Normal file
View File

@ -0,0 +1,155 @@
# Fork Selection
This article describes Solana's *Nakomoto Fork Selection* algorithm based on time
locks. It satisfies the following properties:
* A voter can eventually recover from voting on a fork that doesn't become the
fork with the desired network finality.
* If the voters share a common ancestor then they will converge to a fork
containing that ancestor no matter how they are partitioned. The converged
ancestor may not be the latest possible ancestor at the start of the fork.
* Rollback requires exponentially more time for older votes than for newer
votes.
* Voters have the freedom to set a minimum network confirmation threshold
before committing a vote to a higher lockout. This allows each voter to make
a trade-off between risk and reward. See [cost of rollback](#cost-of-rollback).
## Time
For networks like Solana, time can be the PoH hash count, which is a VDF that
provides a source of time before consensus. Other networks adopting this
approach would need to consider a global source of time.
For Solana, time uniquely identifies a specific leader for fork generation. At
any given time only 1 leader, which can be computed from the ledger itself, can
propose a fork. For more details, see [fork generation](fork-generation.md)
and [leader rotation](leader-rotation.md).
## Algorithm
The basic idea to this approach is to stack consensus votes. Each vote in the
stack is a confirmation of a fork. Each confirmed fork is an ancestor of the
fork above it. Each consensus vote has a `lockout` in units of time before the
validator can submit a vote that does not contain the confirmed fork as an
ancestor.
When a vote is added to the stack, the lockouts of all the previous votes in
the stack are doubled (more on this in [Rollback](#Rollback)). With each new
vote, a voter commits the previous votes to an ever-increasing lockout. At 32
votes we can consider the vote to be at `max lockout` any votes with a lockout
equal to or above `1<<32` are dequeued (FIFO). Dequeuing a vote is the trigger
for a reward. If a vote expires before it is dequeued, it and all the votes
above it are popped (LIFO) from the vote stack. The voter needs to start
rebuilding the stack from that point.
### Rollback
Before a vote is pushed to the stack, all the votes leading up to vote with a
lower lock time than the new vote are popped. After rollback lockouts are not
doubled until the voter catches up to the rollback height of votes.
For example, a vote stack with the following state:
| vote | vote time | lockout | lock expiration time |
|-----:|----------:|--------:|---------------------:|
| 4 | 4 | 2 | 6 |
| 3 | 3 | 4 | 7 |
| 2 | 2 | 8 | 10 |
| 1 | 1 | 16 | 17 |
*Vote 5* is at time 9, and the resulting state is
| vote | vote time | lockout | lock expiration time |
|-----:|----------:|--------:|---------------------:|
| 5 | 9 | 2 | 11 |
| 2 | 2 | 8 | 10 |
| 1 | 1 | 16 | 17 |
*Vote 6* is at time 10
| vote | vote time | lockout | lock expiration time |
|-----:|----------:|--------:|---------------------:|
| 6 | 10 | 2 | 12 |
| 5 | 9 | 4 | 13 |
| 2 | 2 | 8 | 10 |
| 1 | 1 | 16 | 17 |
At time 10 the new votes caught up to the previous votes. But *vote 2* expires
at 10, so the when *vote 7* at time 11 is applied the votes including and above
*vote 2* will be popped.
| vote | vote time | lockout | lock expiration time |
|-----:|----------:|--------:|---------------------:|
| 7 | 11 | 2 | 13 |
| 1 | 1 | 16 | 17 |
The lockout for vote 1 will not increase from 16 until the stack contains 5
votes.
### Slashing and Rewards
The purpose of the lockout is to force a voter to commit opportunity cost to a
specific fork. Voters that violate the lockouts and vote for a diverging fork
within the lockout should be punished. Slashing or simply freezing the voter
from rewards for a long period of time can be used as punishment.
Voters should be rewarded for selecting the fork that the rest of the network
selected as often as possible. This is well-aligned with generating a reward
when the vote stack is full and the oldest vote needs to be dequeued. Thus a
reward should be generated for each successful dequeue.
### Cost of Rollback
Cost of rollback of *fork A* is defined as the cost in terms of lockout time to
the validators to confirm any other fork that does not include *fork A* as an
ancestor.
The **Economic Finality** of *fork A* can be calculated as the loss of all the
rewards from rollback of *fork A* and its descendants, plus the opportunity
cost of reward due to the exponentially growing lockout of the votes that have
confirmed *fork A*.
### Thresholds
Each voter can independently set a threshold of network commitment to a fork
before that voter commits to a fork. For example, at vote stack index 7, the
lockout is 256 time units. A voter may withhold votes and let votes 0-7 expire
unless the vote at index 7 has at greater than 50% commitment in the network.
This allows each voter to independently control how much risk to commit to a
fork. Committing to forks at a higher frequency would allow the voter to earn
more rewards.
### Algorithm parameters
These parameters need to be tuned.
* Number of votes in the stack before dequeue occurs (32).
* Rate of growth for lockouts in the stack (2x).
* Starting default lockout (2).
* Threshold depth for minimum network commitment before committing to the fork
(8).
* Minimum network commitment size at threshold depth (50%+).
### Free Choice
A "Free Choice" is an unenforcible voter action. A voter that maximizes
self-reward over all possible futures should behave in such a way that the
system is stable, and the local greedy choice should result in a greedy choice
over all possible futures. A set of voter that are engaging in choices to
disrupt the protocol should be bound by their stake weight to the denial of
service. Two options exits for voter:
* a voter can outrun previous voters in virtual generation and submit a
concurrent fork
* a voter can withhold a vote to observe multiple forks before voting
In both cases, the voters in the network have several forks to pick from
concurrently, even though each fork represents a different height. In both
cases it is impossible for the protocol to detect if the voter behavior is
intentional or not.
### Greedy Choice for Concurrent Forks
When evaluating multiple forks, each voter should pick the fork that will
maximize economic finality for the network, or the latest fork if all are equal.

598
tests/fork-selection.rs Normal file
View File

@ -0,0 +1,598 @@
//! Fork Selection Simulation
//!
//! Description of the algorithm can be found in [book/src/fork-seleciton.md](book/src/fork-seleciton.md).
//!
//! A test library function exists for configuring networks.
//! ```
//! /// * num_partitions - 1 to 100 partitions
//! /// * fail_rate - 0 to 1.0 rate of packet receive failure
//! /// * delay_count - number of forks to observe before voting
//! /// * parasite_rate - number of parasite nodes that vote oposite the greedy choice
//! fn test_with_partitions(num_partitions: usize, fail_rate: f64, delay_count: usize, parasite_rate: f64);
//! ```
//! Modify the test function
//! ```
//! #[test]
//! #[ignore]
//! fn test_all_partitions() {
//! test_with_partitions(100, 0.0, 5, 0.25, false)
//! }
//! ```
//! Run with cargo
//!
//! ```
//! cargo test all_partitions --release -- --nocapture --ignored
//! ```
//!
//! The output will look like this
//! ```
//! time: 336, tip converged: 76, trunk id: 434, trunk time: 334, trunk converged 98, trunk depth 65
//! ```
//! * time - The current network time. Each packet is transmitted to the network at a different time value.
//! * tip converged - How common is the tip of every voter in the network.
//! * trunk id - fork of every trunk. Every transmission generates a new fork. A trunk is the newest most common fork for the largest converged set of the network.
//! * trunk time - Time when the trunk fork was created.
//! * trunk converged - How many voters have converged on this common fork.
//! * trunk depth - How deep is this fork, or the height of this ledger.
//!
//!
//! ### Simulating Greedy Choice
//!
//! Parasitic nodes reverse the weighted function and pick the fork that has the least amount of economic finality, but without fully committing to a dead fork.
//!
//! ```
//! // Each run starts with 100 partitions, and it takes about 260 forks for a dominant trunk to emerge
//! // fully parasitic, 5 vote delay, 17% efficient
//! test_with_partitions(100, 0.0, 5, 1.0)
//! time: 1000, tip converged: 100, trunk id: 1095, trunk time: 995, trunk converged 100, trunk depth 125
//! // 50% parasitic, 5 vote delay, 30% efficient
//! test_with_partitions(100, 0.0, 5, 0.5)
//! time: 1000, tip converged: 51, trunk id: 1085, trunk time: 985, trunk converged 100, trunk depth 223
//! // 25% parasitic, 5 vote delay, 49% efficient
//! test_with_partitions(100, 0.0, 5, 0.25)
//! time: 1000, tip converged: 79, trunk id: 1096, trunk time: 996, trunk converged 100, trunk depth 367
//! // 0% parasitic, 5 vote delay, 62% efficient
//! test_with_partitions(100, 0.0, 5, 0.0)
//! time: 1000, tip converged: 100, trunk id: 1099, trunk time: 999, trunk converged 100, trunk depth 463
//! // 0% parasitic, 0 vote delay, 100% efficient
//! test_with_partitions(100, 0.0, 0, 0.0)
//! time: 1000, tip converged: 100, trunk id: 1100, trunk time: 1000, trunk converged 100, trunk depth 740
//! ```
//!
//! ### Impact of Receive Errors
//!
//! * with 10% of packet drops, the depth of the trunk is about 77% of the max possible
//! ```
//! time: 4007, tip converged: 94, trunk id: 4005, trunk time: 4002, trunk converged 100, trunk depth 3121
//! ```
//! * with 90% of packet drops, the depth of the trunk is about 8.6% of the max possible
//! ```
//! time: 4007, tip converged: 10, trunk id: 3830, trunk time: 3827, trunk converged 100, trunk depth 348
//! ```
extern crate rand;
use rand::{thread_rng, Rng};
use std::collections::HashMap;
use std::collections::VecDeque;
#[derive(Clone, Default, Debug, Hash, Eq, PartialEq)]
pub struct Fork {
id: usize,
base: usize,
}
impl Fork {
fn is_trunk_of(&self, other: &Fork, fork_tree: &HashMap<usize, Fork>) -> bool {
let mut current = other.clone();
loop {
// found it
if current.id == self.id {
return true;
}
// base is 0, and this id is 0
if current.base == 0 && self.id == 0 {
assert!(fork_tree.get(&0).is_none());
return true;
}
// base is 0
if fork_tree.get(&current.base).is_none() {
return false;
}
current = fork_tree.get(&current.base).unwrap().clone();
}
}
}
#[derive(Clone, Default, Debug, Hash, Eq, PartialEq)]
pub struct Vote {
fork: Fork,
time: usize,
lockout: usize,
}
impl Vote {
pub fn new(fork: Fork, time: usize) -> Vote {
Self {
fork,
time,
lockout: 2,
}
}
pub fn lock_height(&self) -> usize {
self.time + self.lockout
}
pub fn is_trunk_of(&self, other: &Vote, fork_tree: &HashMap<usize, Fork>) -> bool {
self.fork.is_trunk_of(&other.fork, fork_tree)
}
}
#[derive(Debug)]
pub struct LockTower {
votes: VecDeque<Vote>,
max_size: usize,
fork_trunk: Fork,
converge_depth: usize,
delay_count: usize,
delayed_votes: VecDeque<Vote>,
parasite: bool,
}
impl LockTower {
pub fn new(max_size: usize, converge_depth: usize, delay_count: usize) -> Self {
Self {
votes: VecDeque::new(),
max_size,
fork_trunk: Fork::default(),
converge_depth,
delay_count,
delayed_votes: VecDeque::new(),
parasite: false,
}
}
pub fn enter_vote(
&mut self,
vote: Vote,
fork_tree: &HashMap<usize, Fork>,
converge_map: &HashMap<usize, usize>,
scores: &HashMap<Vote, usize>,
) {
let is_valid = self
.get_vote(self.converge_depth)
.map(|v| v.is_trunk_of(&vote, fork_tree))
.unwrap_or(true);
if is_valid {
self.delayed_votes.push_front(vote);
}
loop {
if self.delayed_votes.len() <= self.delay_count {
break;
}
let votes = self.pop_best_votes(fork_tree, scores);
for vote in votes {
self.push_vote(vote, fork_tree, converge_map);
}
}
let trunk = self.votes.get(self.converge_depth).cloned();
trunk.map(|t| {
self.delayed_votes.retain(|v| v.fork.id > t.fork.id);
});
}
pub fn pop_best_votes(
&mut self,
fork_tree: &HashMap<usize, Fork>,
scores: &HashMap<Vote, usize>,
) -> VecDeque<Vote> {
let mut best: Vec<(usize, usize, usize)> = self
.delayed_votes
.iter()
.enumerate()
.map(|(i, v)| (*scores.get(&v).unwrap_or(&0), v.time, i))
.collect();
// highest score, latest vote first
best.sort();
if self.parasite {
best.reverse();
}
// best vote is last
let mut votes: VecDeque<Vote> = best
.last()
.and_then(|v| self.delayed_votes.remove(v.2))
.into_iter()
.collect();
// plus any ancestors
if votes.is_empty() {
return votes;
}
let mut restart = true;
// should really be using heap here
while restart {
restart = false;
for i in 0..self.delayed_votes.len() {
let is_trunk = {
let v = &self.delayed_votes[i];
v.is_trunk_of(votes.front().unwrap(), fork_tree)
};
if is_trunk {
votes.push_front(self.delayed_votes.remove(i).unwrap());
restart = true;
break;
}
}
}
votes
}
pub fn push_vote(
&mut self,
vote: Vote,
fork_tree: &HashMap<usize, Fork>,
converge_map: &HashMap<usize, usize>,
) -> bool {
self.rollback(vote.time);
if !self.is_valid(&vote, fork_tree) {
return false;
}
if !self.is_converged(converge_map) {
return false;
}
self.execute_vote(vote);
if self.is_full() {
self.pop_full();
}
true
}
/// check if the vote at `depth` has over 50% of the network committed
fn is_converged(&self, converge_map: &HashMap<usize, usize>) -> bool {
self.get_vote(self.converge_depth)
.map(|v| {
let v = *converge_map.get(&v.fork.id).unwrap_or(&0);
// hard coded to 100 nodes
assert!(v <= 100);
v > 50
}).unwrap_or(true)
}
pub fn score(&self, vote: &Vote, fork_tree: &HashMap<usize, Fork>) -> usize {
let st = self.rollback_count(vote.time);
if st < self.votes.len() && !self.votes[st].is_trunk_of(vote, fork_tree) {
return 0;
}
let mut rv = 0;
for i in st..self.votes.len() {
let lockout = self.votes[i].lockout;
rv += lockout;
if i == 0 || self.votes[i - 1].lockout * 2 == lockout {
// double the lockout from this vote
rv += lockout;
}
}
rv
}
fn rollback_count(&self, time: usize) -> usize {
let mut last: usize = 0;
for (i, v) in self.votes.iter().enumerate() {
if v.lock_height() < time {
last = i + 1;
}
}
last
}
/// if a vote is expired, pop it and all the votes leading up to it
fn rollback(&mut self, time: usize) {
let last = self.rollback_count(time);
for _ in 0..last {
self.votes.pop_front();
}
}
/// only add votes that are descendent from the last vote in the stack
fn is_valid(&self, vote: &Vote, fork_tree: &HashMap<usize, Fork>) -> bool {
self.last_fork().is_trunk_of(&vote.fork, fork_tree)
}
fn execute_vote(&mut self, vote: Vote) {
let vote_time = vote.time;
assert!(!self.is_full());
assert_eq!(vote.lockout, 2);
// push the new vote to the font
self.votes.push_front(vote);
// double the lockouts if the threshold to doulbe is met
for i in 1..self.votes.len() {
assert!(self.votes[i].time <= vote_time);
if self.votes[i].lockout == self.votes[i - 1].lockout {
self.votes[i].lockout *= 2;
}
}
}
fn pop_full(&mut self) {
assert!(self.is_full());
self.fork_trunk = self.votes.pop_back().unwrap().fork;
}
fn is_full(&self) -> bool {
assert!(self.votes.len() <= self.max_size);
self.votes.len() == self.max_size
}
fn last_vote(&self) -> Option<&Vote> {
self.votes.front()
}
fn get_vote(&self, ix: usize) -> Option<&Vote> {
self.votes.get(ix)
}
pub fn first_vote(&self) -> Option<&Vote> {
self.votes.back()
}
pub fn last_fork(&self) -> Fork {
self.last_vote()
.map(|v| v.fork.clone())
.unwrap_or(self.fork_trunk.clone())
}
}
#[test]
fn test_is_trunk_of_1() {
let tree = HashMap::new();
let b1 = Fork { id: 1, base: 0 };
let b2 = Fork { id: 2, base: 0 };
assert!(!b1.is_trunk_of(&b2, &tree));
}
#[test]
fn test_is_trunk_of_2() {
let tree = HashMap::new();
let b1 = Fork { id: 1, base: 0 };
let b2 = Fork { id: 0, base: 0 };
assert!(!b1.is_trunk_of(&b2, &tree));
}
#[test]
fn test_is_trunk_of_3() {
let tree = HashMap::new();
let b1 = Fork { id: 1, base: 0 };
let b2 = Fork { id: 1, base: 0 };
assert!(b1.is_trunk_of(&b2, &tree));
}
#[test]
fn test_is_trunk_of_4() {
let mut tree = HashMap::new();
let b1 = Fork { id: 1, base: 0 };
let b2 = Fork { id: 2, base: 1 };
tree.insert(b1.id, b1.clone());
assert!(b1.is_trunk_of(&b2, &tree));
assert!(!b2.is_trunk_of(&b1, &tree));
}
#[test]
fn test_push_vote() {
let tree = HashMap::new();
let bmap = HashMap::new();
let b0 = Fork { id: 0, base: 0 };
let mut node = LockTower::new(32, 7, 0);
let vote = Vote::new(b0.clone(), 0);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes.len(), 1);
let vote = Vote::new(b0.clone(), 1);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes.len(), 2);
let vote = Vote::new(b0.clone(), 2);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes.len(), 3);
let vote = Vote::new(b0.clone(), 3);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes.len(), 4);
assert_eq!(node.votes[0].lockout, 2);
assert_eq!(node.votes[1].lockout, 4);
assert_eq!(node.votes[2].lockout, 8);
assert_eq!(node.votes[3].lockout, 16);
assert_eq!(node.votes[1].lock_height(), 6);
assert_eq!(node.votes[2].lock_height(), 9);
let vote = Vote::new(b0.clone(), 7);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes[0].lockout, 2);
let b1 = Fork { id: 1, base: 1 };
let vote = Vote::new(b1.clone(), 8);
assert!(!node.push_vote(vote, &tree, &bmap));
let vote = Vote::new(b0.clone(), 8);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes.len(), 4);
assert_eq!(node.votes[0].lockout, 2);
assert_eq!(node.votes[1].lockout, 4);
assert_eq!(node.votes[2].lockout, 8);
assert_eq!(node.votes[3].lockout, 16);
let vote = Vote::new(b0.clone(), 10);
assert!(node.push_vote(vote, &tree, &bmap));
assert_eq!(node.votes.len(), 2);
assert_eq!(node.votes[0].lockout, 2);
assert_eq!(node.votes[1].lockout, 16);
}
fn create_network(sz: usize, depth: usize, delay_count: usize) -> Vec<LockTower> {
(0..sz)
.into_iter()
.map(|_| LockTower::new(32, depth, delay_count))
.collect()
}
/// The "height" or "depth" of this fork. How many forks until it connects to fork 0
fn calc_fork_depth(fork_tree: &HashMap<usize, Fork>, id: usize) -> usize {
let mut depth = 0;
let mut start = fork_tree.get(&id);
loop {
if start.is_none() {
break;
}
depth += 1;
start = fork_tree.get(&start.unwrap().base);
}
depth
}
/// map of `fork id` to `node count`
/// This map contains how many nodes have the fork as an ancestor
/// The fork with the highest count that is the newest is the network "trunk"
fn calc_fork_map(
network: &Vec<LockTower>,
fork_tree: &HashMap<usize, Fork>,
) -> HashMap<usize, usize> {
let mut lca_map: HashMap<usize, usize> = HashMap::new();
for node in network {
let mut start = node.last_fork();
loop {
*lca_map.entry(start.id).or_insert(0) += 1;
if fork_tree.get(&start.base).is_none() {
break;
}
start = fork_tree.get(&start.base).unwrap().clone();
}
}
lca_map
}
/// find the fork with the highest count of nodes that have it as an ancestor
/// as well as with the highest possible fork id, which indicates it is the newest
fn calc_newest_trunk(bmap: &HashMap<usize, usize>) -> (usize, usize) {
let mut data: Vec<_> = bmap.iter().collect();
data.sort_by_key(|x| (x.1, x.0));
data.last().map(|v| (*v.0, *v.1)).unwrap()
}
/// how common is the latest fork of all the nodes
fn calc_tip_converged(network: &Vec<LockTower>, bmap: &HashMap<usize, usize>) -> usize {
let sum: usize = network
.iter()
.map(|n| *bmap.get(&n.last_fork().id).unwrap_or(&0))
.sum();
sum / network.len()
}
#[test]
fn test_no_partitions() {
let mut tree = HashMap::new();
let len = 100;
let mut network = create_network(len, 32, 0);
for rounds in 0..1 {
for i in 0..network.len() {
let time = rounds * len + i;
let base = network[i].last_fork().clone();
let fork = Fork {
id: time + 1,
base: base.id,
};
tree.insert(fork.id, fork.clone());
let vote = Vote::new(fork, time);
let bmap = calc_fork_map(&network, &tree);
for node in network.iter_mut() {
assert!(node.push_vote(vote.clone(), &tree, &bmap));
}
println!("{} {}", time, calc_tip_converged(&network, &bmap));
}
}
let bmap = calc_fork_map(&network, &tree);
assert_eq!(calc_tip_converged(&network, &bmap), len);
}
/// * num_partitions - 1 to 100 partitions
/// * fail_rate - 0 to 1.0 rate of packet receive failure
/// * delay_count - number of forks to observe before voting
/// * parasite_rate - number of parasite nodes that vote oposite the greedy choice
fn test_with_partitions(
num_partitions: usize,
fail_rate: f64,
delay_count: usize,
parasite_rate: f64,
break_early: bool,
) {
let mut fork_tree = HashMap::new();
let len = 100;
let warmup = 8;
let mut network = create_network(len, warmup, delay_count);
for time in 0..warmup {
let bmap = calc_fork_map(&network, &fork_tree);
for node in network.iter_mut() {
let mut fork = node.last_fork().clone();
if fork.id == 0 {
fork.id = thread_rng().gen_range(1, 1 + num_partitions);
fork_tree.insert(fork.id, fork.clone());
}
let vote = Vote::new(fork, time);
assert!(node.is_valid(&vote, &fork_tree));
assert!(node.push_vote(vote.clone(), &fork_tree, &bmap));
}
}
for node in network.iter_mut() {
assert_eq!(node.votes.len(), warmup);
assert_eq!(node.first_vote().unwrap().lockout, 1 << warmup);
assert!(node.first_vote().unwrap().lock_height() >= 1 << warmup);
node.parasite = parasite_rate > thread_rng().gen_range(0.0, 1.0);
}
let converge_map = calc_fork_map(&network, &fork_tree);
assert_ne!(calc_tip_converged(&network, &converge_map), len);
for rounds in 0..10 {
for i in 0..len {
let time = warmup + rounds * len + i;
let base = network[i].last_fork().clone();
let fork = Fork {
id: time + num_partitions,
base: base.id,
};
fork_tree.insert(fork.id, fork.clone());
let converge_map = calc_fork_map(&network, &fork_tree);
let vote = Vote::new(fork, time);
let mut scores: HashMap<Vote, usize> = HashMap::new();
network.iter().for_each(|n| {
n.delayed_votes.iter().for_each(|v| {
*scores.entry(v.clone()).or_insert(0) += n.score(&v, &fork_tree);
})
});
for node in network.iter_mut() {
if thread_rng().gen_range(0f64, 1.0f64) < fail_rate {
continue;
}
node.enter_vote(vote.clone(), &fork_tree, &converge_map, &scores);
}
let converge_map = calc_fork_map(&network, &fork_tree);
let trunk = calc_newest_trunk(&converge_map);
let trunk_time = if trunk.0 > num_partitions {
trunk.0 - num_partitions
} else {
trunk.0
};
println!(
"time: {}, tip converged: {}, trunk id: {}, trunk time: {}, trunk converged {}, trunk depth {}",
time,
calc_tip_converged(&network, &converge_map),
trunk.0,
trunk_time,
trunk.1,
calc_fork_depth(&fork_tree, trunk.0)
);
if break_early && calc_tip_converged(&network, &converge_map) == len {
break;
}
}
if break_early {
let converge_map = calc_fork_map(&network, &fork_tree);
if calc_tip_converged(&network, &converge_map) == len {
break;
}
}
}
let converge_map = calc_fork_map(&network, &fork_tree);
let trunk = calc_newest_trunk(&converge_map);
assert_eq!(trunk.1, len);
}
#[test]
fn test_3_partitions() {
test_with_partitions(3, 0.0, 0, 0.0, true)
}
#[test]
#[ignore]
fn test_3_partitions_large_packet_drop() {
test_with_partitions(3, 0.9, 0, 0.0, false)
}
#[test]
#[ignore]
fn test_all_partitions() {
test_with_partitions(100, 0.0, 5, 0.25, false)
}