Skip to main content
networking

Latency

Latency is the time delay between a request being sent and the first byte of a response being received. Measured in milliseconds. Network latency is dominated by the speed of light over physical distance (1ms per 100km), plus queuing and processing delays.

Definition

Network latency is the round-trip time (RTT) from sender to receiver and back. It has a hard physical lower bound: light travels approximately 200km/ms in fiber (accounting for refraction). A request from New York to London (~5,500km) has a theoretical minimum of ~27ms RTT. Real-world latency adds: TCP handshake overhead, TLS negotiation, DNS resolution, server processing, and network queuing. Latency is distinct from bandwidth – a satellite link may have high bandwidth but very high latency (~600ms). Latency is the primary performance constraint for interactive applications, while bandwidth matters more for bulk transfers.

Examples

  • CDN reduces latency by serving from edge nodes geographically near users
  • TCP has 1 RTT handshake overhead before data can flow
  • TLS 1.3 adds 1 RTT for the TLS handshake
  • 0-RTT resumption in TLS 1.3 / QUIC eliminates handshake latency for returning clients

Related Protocols

Related Terms