QoS

Quality of Service (QoS) is a set of techniques used in computer networks to manage resource allocation and ensure predictable performance for selected types of traffic. QoS aims to provide sufficient bandwidth, lower latency, reduced jitter, and/or lower packet loss for time-sensitive or high-priority applications while allowing other traffic to use remaining capacity.

QoS mechanisms typically cover classification, marking, policing, shaping, scheduling, and congestion management. Classification assigns packets to traffic classes, often using markings such as DSCP in IP headers or 802.1p in Ethernet. Marking guides how routers treat packets as they pass through a network. Scheduling and queue management determine transmission order and rate, using methods like priority queuing, weighted fair queuing, and deficit round robin, along with active queue management to prevent bufferbloat. Policing and shaping control traffic rates during ingress and egress. At the architecture level, QoS can be implemented with Integrated Services (IntServ), which uses explicit resource reservations (RSVP) for individual flows, or Differentiated Services (DiffServ), which handles traffic classes at hop boundaries. MPLS can carry QoS information across an entire path to support traffic engineering.

QoS targets are common for voice over IP, video conferencing, real-time gaming, and other latency‑sensitive applications, as well as business-critical services. Key metrics include available bandwidth, end-to-end latency, jitter, and packet loss. QoS effectiveness depends on network design and policy, and it is most reliable within controlled domains; end-to-end guarantees across the public Internet are limited by varying support and encryption. Wireless networks may use standards like IEEE 802.11e/WMM to prioritize traffic.