What is Quality of Service in networking, and why is it necessary? Explain the key metrics that QoS addresses: latency, jitter, packet loss, and bandwidth.

Question

Accepted Answer

Quality of Service is a set of technologies and techniques that manage network traffic to ensure critical applications receive the performance they need. Without QoS, all traffic is treated equally on a best-effort basis. When network links become congested, packets are dropped indiscriminately, and delay-sensitive applications like voice and video suffer alongside bulk data transfers. QoS provides the mechanisms to prioritize important traffic, guarantee minimum bandwidth, and control how congestion affects different traffic types.

Latency, also called delay, is the time it takes for a packet to travel from source to destination. It is measured in milliseconds and includes serialization delay, propagation delay, queuing delay, and processing delay at each hop. Voice and video applications are particularly sensitive to latency because users experience it as perceptible delay in conversation. For voice over IP, one-way latency should remain below 150 milliseconds for acceptable call quality, with the ITU G.114 recommendation considering anything above this as potentially impaired. Interactive applications like remote desktop and video conferencing also require low latency for a responsive user experience.

Jitter is the variation in latency between consecutive packets. Even if average latency is acceptable, high jitter means some packets arrive quickly while others are significantly delayed. Real-time applications use buffers to smooth out jitter, but the buffer adds additional latency. If jitter exceeds the buffer capacity, packets arrive too late to be useful and are effectively lost. For voice traffic, jitter should ideally remain below 30 milliseconds. Consistent latency with low jitter is often more important than having the absolute lowest average latency.

Packet loss occurs when packets are dropped somewhere in the network, most commonly at congested router or switch queues that overflow. TCP-based applications handle packet loss through retransmission, which increases latency but maintains data integrity. UDP-based real-time applications like voice and video cannot wait for retransmissions, so lost packets result in gaps in audio or video. For voice, packet loss above one percent begins to noticeably degrade quality. For video, the impact depends on the codec and whether the lost packet contained a key frame.

Bandwidth refers to the maximum data rate of a network link. While raw bandwidth is a capacity measurement, QoS manages how that bandwidth is allocated among competing traffic types. Without QoS, a single large file transfer can consume all available bandwidth on a link, starving voice and video traffic. QoS bandwidth management reserves a minimum portion of link capacity for critical traffic classes, ensures no single traffic type can monopolize the link, and limits lower-priority traffic during congestion.

QoS is necessary because modern enterprise networks carry diverse traffic types with different performance requirements over shared infrastructure. Voice, video, interactive applications, bulk data, and background traffic all share the same network links, and the network must balance their competing needs. QoS becomes critical at congestion points where demand exceeds capacity, which can occur on WAN links, internet connections, wireless uplinks, and even campus network links during peak usage.