Ping / Latency Estimator

Estimates the theoretical minimum round-trip time between two locations based on distance, medium, and hop count — useful for setting realistic expectations before a VoIP rollout, a distributed system design, or an SLA negotiation with a site on the other side of the planet.

The physics you can't engineer around

Signal travels through fibre at roughly 200,000 km/s — about two-thirds the speed of light in a vacuum. That's a hard physical floor, not an engineering limitation you can buy your way past. London to New York, roughly 5,500km, works out to a minimum of about 27ms one-way — call it 55ms round trip, before routing, processing delay or congestion add anything on top. No amount of hardware spend changes that number; it's set by distance and the speed of light, full stop.

Latency by medium

• Fibre / copper Ethernet — lowest latency for the distance, roughly equivalent to each other over short-to-medium runs
• GEO satellite — ~35,786km up, minimum 600ms round trip. Not usable for anything real-time.
• LEO satellite (Starlink) — ~550km up, 20–60ms round trip — genuinely usable for most applications, unlike GEO
• 4G/LTE — 30–100ms typical, fine for most business use
• 5G — down to single-digit milliseconds in the best case, which is what makes edge computing use cases actually viable

Why this matters more than people assume

A web page firing 50 sequential database queries over a 100ms-latency connection takes a minimum of 5 seconds to render, regardless of how fast each individual query executes — which is exactly why application and database servers sit on the same LAN with sub-millisecond latency between them, not across a WAN link. The same logic applies to microservices: latency between services compounds across every hop in a call chain, which is a big part of why service topology and colocation decisions matter as much as raw bandwidth.