DNS Benchmark: Cached vs Non-Cached Performance

• cached.py → Measures cached DNS performance
• non-cached.py → Measures uncached DNS performance using random domains

Requirements

• Python 3.x
• dig command available (install via dnsutils or bind-tools)

Example Usage

$ python cached.py

=== CACHED DNS TEST ===
Domain: google.com
Runs per server: 20


Testing Cloudflare (1.1.1.1)
[20/20] Querying: google.com

Testing Google (8.8.8.8)
[20/20] Querying: google.com

Testing Quad9 (9.9.9.9)
[20/20] Querying: google.com

=== CACHED DNS RESULTS ===

Cloudflare
  Median: 5.00 ms
  Avg:    6.10 ms
  P95:    19.00 ms

Google
  Median: 4.00 ms
  Avg:    5.65 ms
  P95:    18.00 ms

Quad9
  Median: 4.00 ms
  Avg:    5.65 ms
  P95:    20.00 ms

Metrics Explanation

Each script outputs:

• Median (P50) → Typical response time (what users usually experience)
• Average (Avg) → Overall performance, but can be misleading due to spikes
• P95 → 95% of queries are faster than this value. Reveals latency spikes and instability

👉 Key idea: Average shows speed. Median shows reality. P95 shows pain.

Recommended Settings

You can modify these values inside the scripts.

Default: RUNS = 20

Suggested:

• Minimum: 20
• Ideal: 30–50

Adding More DNS Providers

Edit this section in both scripts:

SERVERS = { "Cloudflare": "1.1.1.1", "Google": "8.8.8.8", "Quad9": "9.9.9.9" }

Example additions:

• "OpenDNS": "208.67.222.222",
• "AdGuard": "94.140.14.14"

License

Licensed under MIT. Use freely and modify aggressively.