Latency refers to the time interval between a user action and the resulting response from the system. In networking, it is often measured as Round-Trip Time (RTT), representing the duration for a data packet to travel to its destination and back to the sender. It is a fundamental metric for evaluating network responsiveness and performance.
The metric is influenced by physical distance, the medium of transmission (fiber optics vs. copper), and the number of network hops. Key components include propagation delay, processing delay at routers, and queuing delay during periods of high network congestion. Unlike bandwidth, which measures data volume over time, latency measures the speed of a single interaction.
Latency is a critical performance indicator for real-time services such as VoIP, online gaming, and financial trading platforms. High latency can cause synchronization issues in distributed systems and microservices architectures, where multiple internal requests must complete before a response is sent to the client.
Measurement and Standards
Typical latency for a fiber-optic connection within a metropolitan area is under 10ms. For intercontinental traffic, the physical limits of light speed dictate a minimum latency of approximately 65ms between London and New York. In high-frequency trading (HFT), even a 1ms delay can result in significant financial consequences, necessitating specialized low-latency hardware.