auradb 0.1.0

High-performance Rust storage engine with WAL-time KV separation, RL-driven compaction, and learned indexes
Documentation 
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
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206

use std::time::Instant;
use std::fs;
use std::path::Path;
use auradb::AuraEngine;
use auradb::config::Config;
use rand::Rng;

fn run_workload(label: &str, operations: usize, key_size: usize, value_size: usize, read_ratio: f64) {
    println!("\n📊 Running workload: {label}");
    println!("   Operations: {} | Key size: {}B | Value size: {}B | Read ratio: {:.0}%", 
             operations, key_size, value_size, read_ratio * 100.0);
    
    let config = Config::default();
    let mut db = AuraEngine::new(config).expect("DB init failed");

    let mut rng = rand::thread_rng();
    let mut keys: Vec<Vec<u8>> = Vec::with_capacity(operations);

    let start = Instant::now();

    for i in 0..operations {
        let key: Vec<u8> = (0..key_size).map(|_| rng.gen::<u8>()).collect();
        let val: Vec<u8> = (0..value_size).map(|_| rng.gen::<u8>()).collect();

        if rng.gen::<f64>() < read_ratio && !keys.is_empty() {
            let k = &keys[rng.gen_range(0..keys.len())];
            let _ = db.get_bytes(k);
        } else {
            db.put_bytes(&key, &val).expect("Put failed");
            keys.push(key);
        }
        
        // Progress indicator for long-running workloads
        if i % (operations / 10).max(1) == 0 {
            print!("\r   Progress: {}/{} ({:.0}%)", i, operations, (i as f64 / operations as f64) * 100.0);
        }
    }
    println!(); // Clear progress line

    let duration = start.elapsed().as_secs_f64();

    let throughput_ops = operations as f64 / duration;
    let throughput_mb = (operations as f64 * value_size as f64) / (1024.0 * 1024.0 * duration);

    println!("✅ Duration: {:.2} s", duration);
    println!("🚀 Throughput: {:.0} ops/sec | {:.2} MB/s", throughput_ops, throughput_mb);
    println!("💾 Disk size: {:.2} MB", folder_size("./benchmark_db") as f64 / (1024.0 * 1024.0));
}

fn run_latency_workload(label: &str, operations: usize, key_size: usize, value_size: usize) {
    println!("\n⏱ Running latency workload: {label}");
    println!("   Operations: {} | Key size: {}B | Value size: {}B", operations, key_size, value_size);
    
    let config = Config::default();
    let db = AuraEngine::new(config).expect("DB init failed");

    let mut rng = rand::thread_rng();
    let mut latencies = Vec::with_capacity(operations);

    // Pre-populate with some data
    println!("   📝 Pre-populating database...");
    for i in 0..(operations / 10) {
        let key: Vec<u8> = (0..key_size).map(|_| rng.gen::<u8>()).collect();
        let val: Vec<u8> = (0..value_size).map(|_| rng.gen::<u8>()).collect();
        db.put_bytes(&key, &val).expect("Put failed");
    }

    println!("   🔍 Measuring read latencies...");
    for i in 0..operations {
        let key: Vec<u8> = (0..key_size).map(|_| rng.gen::<u8>()).collect();
        
        let start = Instant::now();
        let _ = db.get_bytes(&key);
        let elapsed = start.elapsed().as_nanos() as u64;
        latencies.push(elapsed);
        
        if i % (operations / 10).max(1) == 0 {
            print!("\r   Progress: {}/{} ({:.0}%)", i, operations, (i as f64 / operations as f64) * 100.0);
        }
    }
    println!(); // Clear progress line

    // Calculate percentiles
    latencies.sort();
    let p50 = latencies[latencies.len() * 50 / 100];
    let p95 = latencies[latencies.len() * 95 / 100];
    let p99 = latencies[latencies.len() * 99 / 100];
    let p999 = latencies[latencies.len() * 999 / 1000];

    println!("✅ Latency percentiles:");
    println!("   p50:  {} ns ({:.2} µs)", p50, p50 as f64 / 1000.0);
    println!("   p95:  {} ns ({:.2} µs)", p95, p95 as f64 / 1000.0);
    println!("   p99:  {} ns ({:.2} µs)", p99, p99 as f64 / 1000.0);
    println!("   p99.9: {} ns ({:.2} µs)", p999, p999 as f64 / 1000.0);
}

fn run_mixed_workload(label: &str, operations: usize, key_size: usize, value_size: usize, write_ratio: f64) {
    println!("\n🔀 Running mixed workload: {label}");
    println!("   Operations: {} | Key size: {}B | Value size: {}B | Write ratio: {:.0}%", 
             operations, key_size, value_size, write_ratio * 100.0);
    
    let config = Config::default();
    let db = AuraEngine::new(config).expect("DB init failed");

    let mut rng = rand::thread_rng();
    let mut keys: Vec<Vec<u8>> = Vec::new();
    let mut write_count = 0;
    let mut read_count = 0;

    let start = Instant::now();

    for i in 0..operations {
        if rng.gen::<f64>() < write_ratio || keys.is_empty() {
            // Write operation
            let key: Vec<u8> = (0..key_size).map(|_| rng.gen::<u8>()).collect();
            let val: Vec<u8> = (0..value_size).map(|_| rng.gen::<u8>()).collect();
            db.put_bytes(&key, &val).expect("Put failed");
            keys.push(key);
            write_count += 1;
        } else {
            // Read operation
            let k = &keys[rng.gen_range(0..keys.len())];
            let _ = db.get_bytes(k);
            read_count += 1;
        }
        
        if i % (operations / 10).max(1) == 0 {
            print!("\r   Progress: {}/{} ({:.0}%)", i, operations, (i as f64 / operations as f64) * 100.0);
        }
    }
    println!(); // Clear progress line

    let duration = start.elapsed().as_secs_f64();

    let total_throughput = operations as f64 / duration;
    let write_throughput = write_count as f64 / duration;
    let read_throughput = read_count as f64 / duration;

    println!("✅ Duration: {:.2} s", duration);
    println!("🚀 Total throughput: {:.0} ops/sec", total_throughput);
    println!("📝 Write throughput: {:.0} ops/sec ({:.0}%)", write_throughput, (write_count as f64 / operations as f64) * 100.0);
    println!("📖 Read throughput: {:.0} ops/sec ({:.0}%)", read_throughput, (read_count as f64 / operations as f64) * 100.0);
}

fn folder_size<P: AsRef<Path>>(path: P) -> u64 {
    match fs::read_dir(path) {
        Ok(entries) => {
            entries.filter_map(|f| {
                f.ok().and_then(|entry| {
                    if let Ok(file_type) = entry.file_type() {
                        if file_type.is_dir() {
                            Some(folder_size(entry.path()))
                        } else {
                            entry.metadata().ok().map(|m| m.len())
                        }
                    } else {
                        None
                    }
                })
            }).sum()
        }
        Err(_) => 0, // Return 0 if directory doesn't exist or can't be read
    }
}

fn main() {
    println!("🚀 Full AuraDB Benchmark Suite");
    println!("================================");
    
    // Clean up any existing database
    fs::remove_dir_all("./benchmark_db").ok();
    
    // Test 1: Small KV writes (1 KB) — measures raw ops/sec
    run_workload("Small Writes (1 KB)", 100_000, 16, 1024, 0.0);
    fs::remove_dir_all("./benchmark_db").ok();

    // Test 2: Medium KV writes (8 KB) — measures MB/s scaling
    run_workload("Medium Writes (8 KB)", 50_000, 16, 8192, 0.0);
    fs::remove_dir_all("./benchmark_db").ok();

    // Test 3: Large KV writes (64 KB) — tests WAL-time KV separation benefit
    run_workload("Large Writes (64 KB)", 10_000, 16, 65536, 0.0);
    fs::remove_dir_all("./benchmark_db").ok();

    // Test 4: Read-heavy workload (80% reads, 20% writes) — tests cache efficiency
    run_mixed_workload("Read Heavy (80% reads)", 50_000, 16, 1024, 0.2);
    fs::remove_dir_all("./benchmark_db").ok();

    // Test 5: Write-heavy workload (80% writes, 20% reads)
    run_mixed_workload("Write Heavy (80% writes)", 50_000, 16, 1024, 0.8);
    fs::remove_dir_all("./benchmark_db").ok();

    // Test 6: Latency profiling for small values
    run_latency_workload("Latency Profile (1 KB)", 100_000, 16, 1024);
    fs::remove_dir_all("./benchmark_db").ok();

    println!("\n🎯 Benchmark Summary");
    println!("====================");
    println!("✅ All workloads completed successfully!");
    println!("📊 Check the results above to identify:");
    println!("   • Raw speed bottlenecks (CPU vs I/O)");
    println!("   • Value size impact on performance");
    println!("   • Read/write ratio effects");
    println!("   • Latency distribution characteristics");
    println!("   • Expected WAL-time KV separation benefits for large values");
}