Mastering QoS: Ensure Reliable Voice, Video, and Data Performance

QoS (Quality of Service) allocates limited bandwidth among different services to provide end‑to‑end performance guarantees, ensuring that voice, video, and critical data applications receive priority handling in network devices.

Importance of QoS

IP network traffic includes real‑time services (e.g., voice) that require stable bandwidth and low latency, and non‑real‑time services with unpredictable bandwidth usage that can cause congestion, packet loss, and degraded quality.

The most effective way to alleviate congestion, aside from adding bandwidth, is to apply a guaranteed‑service strategy that manages traffic flow.

When sustained traffic exceeds bandwidth limits, network expansion or dedicated equipment may be needed to maintain service quality.

Rapid growth of high‑resolution video on smartphones and the rise of video conferencing, surveillance, and streaming generate massive, often unstable video traffic, increasing the challenges for QoS design.

QoS Measurement Metrics

Key metrics include bandwidth, latency, jitter, and packet‑loss rate.

1) Bandwidth

Bandwidth (throughput) is the maximum number of bits transferred per second between two network points, measured in bit/s. It includes upstream (user‑to‑network) and downstream (network‑to‑user) rates.

2) Latency

Latency is the delay from sending to receiving a packet, comprising transmission and processing delays. Delays under 100 ms are generally imperceptible; 100‑300 ms cause noticeable pauses, and >300 ms degrade conversational quality.

3) Jitter

Jitter describes variation in packet delay, calculated as the difference between maximum and minimum latency. High jitter disrupts real‑time audio/video and can affect protocol behavior.

4) Packet‑Loss Rate

Packet loss is the percentage of packets that fail to reach the destination. Small loss may be tolerable for voice, but larger loss degrades video and overall transmission efficiency.

QoS Application Scenarios

In enterprise environments, QoS can prioritize Telnet, video conferencing, VoIP, FTP, and video playback while leaving less critical traffic at default settings.

(1) Network and Management Protocols (e.g., OSPF, Telnet) – Require low latency and loss but modest bandwidth; QoS can assign higher priority.

(2) Real‑time Services (e.g., video conferencing, VoIP) – Demand high bandwidth, low latency, low jitter; QoS can allocate sufficient bandwidth and raise priority.

(3) High‑Volume Data (e.g., FTP, database backup) – Need minimal packet loss; QoS can use traffic shaping to buffer bursts.

(4) Streaming Media (e.g., online audio/video) – Tolerate some buffering; QoS can raise priority to reduce loss and delay.

(5) Ordinary Traffic (e.g., web browsing, email) – No special QoS requirements; default settings are sufficient.

QoS Service Models

QoS models define how end‑to‑end quality guarantees are achieved across network devices.

(1) Best‑Effort Model

Network forwards packets without guaranteeing latency or loss; suitable for non‑critical services like FTP and email.

(2) IntServ Model

Applications signal their traffic parameters via RSVP, and the network reserves resources to meet those guarantees, maintaining per‑flow state.

(3) DiffServ Model

Traffic is classified into classes at the network edge, marked with DSCP or similar fields, and routers treat each class differently without per‑flow state, providing scalable QoS.

Components Based on DiffServ

(1) Packet Classification and Marking – Packets are categorized and assigned priority values (e.g., VLAN 802.1p, IP DSCP, MPLS EXP).

(2) Traffic Policing, Shaping, and Interface Rate Limiting – Policing drops excess traffic; shaping buffers it; rate limiting controls interface bandwidth.

(3) Congestion Management and Avoidance – Queues excess packets and schedules forwarding; avoidance proactively drops packets to prevent overload.

These functions are applied in the order shown in the diagram below.

Comparison of QoS and HQoS

Traditional QoS, based on port‑level scheduling, works well in core networks but cannot differentiate users or manage multiple services per user, limiting its suitability for access networks.

Hierarchical QoS (HQoS) introduces multi‑level queues to distinguish users and services, enabling fine‑grained, cost‑effective resource control.

Source: https://support.huawei.com/enterprise/en/doc/EDOC1100219443