Overview

In computing and networking, bandwidth is commonly used to describe the maximum amount of digital information that can be transmitted from one point to another over a communication channel in a given period of time. In practice it is expressed as a bit rate measured in bits per second (for example Mbps). The term appears frequently in contexts such as computer networks and computer science when discussing capacity, planning, or performance. It is also used informally to refer to available capacity for an application or user.

Characteristics and units

Bandwidth is a theoretical or advertised capacity: the rate a link or device can carry data under ideal conditions. Common units scale by thousands (kbps, Mbps, Gbps). Practical delivered rates—often called throughput—are usually lower than nominal bandwidth because of protocol overhead, congestion, and interference. Important related metrics include latency (delay) and jitter (variability in delay), which affect perceived performance even when bandwidth is sufficient.

History and theoretical limits

The concept stems from signal processing where bandwidth referred to a range of frequencies. In digital communications the idea evolved to capacity limits. The Shannon–Hartley theorem from information theory links channel bandwidth, signal power and noise to an upper bound on error-free data rate; it provides a theoretical ceiling but does not predict actual throughput in a networked system.

Uses and examples

Bandwidth considerations guide the design and use of many systems: streaming video requires sustained bit rates, file transfer benefits from high peak capacity, and interactive services (voice, gaming, remote desktop) are sensitive to both bandwidth and latency. Service providers advertise bandwidth tiers; administrators plan capacity to avoid congestion and ensure quality of service.

Measurement, management and distinctions

Tools measure offered bandwidth and observed throughput to diagnose problems. Management techniques include traffic shaping, prioritization (QoS), compression, and caching. It is important to distinguish between related terms:

  • Bandwidth — channel or link capacity under ideal conditions.
  • Throughput — actual observed data rate.
  • Latency — time for a packet to travel end to end.
  • Spectral bandwidth — frequency range in analog signals, historically related but different in meaning.

For further reading on practical measurements and protocol effects, see material on data transmission, network engineering basics, and tools for bandwidth testing. Vendor documentation and standards notes often explain how advertised rates are defined and measured. Additional resources are available through general networking and computer science references bits, seconds, and performance testing guides bit rate and network textbooks.