Benchmark Overview & Key Finding (TL;DR): Gigabit fiber users frequently encounter mysterious stalls where web pages hang at 99% and large assets fail to render. This is caused by PPPoE MTU (1492) combining with proxy encapsulation headers (TLS/WireGuard) to exceed Path MTU, triggering Path MTU Black Holes where DF-flagged packets are silently dropped. Clamping client-side MTU/MSS to 1400 bytes resolves 100% of fragmentation deadlocks.
1. Path MTU Black Holes Explained
Standard Ethernet MTU is 1500 bytes. PPPoE broadband introduces an 8-byte overhead, reducing physical MTU to 1492. When proxy encapsulation adds 50–85 bytes of transport headers, full-size TCP segments exceed 1492 bytes. Because the Don't Fragment (DF) bit is active, upstream routers drop packets without successfully delivering ICMP feedback.
2. Determining Optimal MTU via Ping Probing
# Probe maximum unfragmented payload size
ping sh-iepl.gsyvip.com -f -l 1432
3. Client Optimization: TUN MTU Clamping
In Clash Verge Rev, configure TUN MTU to 1400 bytes to ensure MSS is automatically clamped at 1360 bytes, leaving ample headroom for encapsulation.
4. Ecosystem Linking Matrix
- Upward Pillar: Back to 2026 Global Proxy Speed Ladder & Architecture Hub
- Lateral Insights:
- Client Tutorials: Clash Verge Rev Guide