The OcNOS QoS Guide provides a comprehensive reference to the QoS framework that enables deterministic packet forwarding and resource allocation across network interfaces. By prioritizing critical data streams over standard traffic during network contention, QoS prevents packet loss for latency-sensitive applications and ensures optimal bandwidth utilization.

QoS prioritizes certain traffic over others, improving network performance and bandwidth utilization. Without QoS, all traffic has the same chance of being delivered or dropped during congestion.

This guide is intended for:

Network Administrators and Operations Engineers responsible for deploying, configuring, and monitoring daily QoS policies, interface queues, rate-limits, and traffic statistics via CLI or NetConf.

Network Architects designing end-to-end traffic management architectures, DiffServ/CoS mapping models, bandwidth allocation strategies, and SLA compliance frameworks.

Service Provider and Carrier Engineers configuring Customer Edge (CE) to Provider Edge (PE) mappings, Attachment Circuits, or Hierarchical QoS (HQoS/hierarchical policing).

Data Center & Cloud Infrastructure Engineers managing multi-tenant traffic isolation, congestion control (ECN, WRED), and lossless fabric transport (PFC/RoCEv2).

This guide enables users to perform the following:

Granular Traffic Classification using Access Control Lists (ACLs), Layer 2 fields (VLAN, CoS, EtherType), and Layer 3 headers (IPv4/IPv6 DSCP, IP Precedence).

Traffic Policing & Metering: Ingress rate limiting via Committed Information Rate (CIR) and Excess Information Rate (EIR); supports single-rate/two-rate policing and Hierarchical Traffic Policing per attachment circuit and traffic class.

Packet Marking & Remarking of ingress and egress traffic to preserve QoS policy across network boundaries.

Flexible Egress Scheduling ensuring latency-sensitive traffic is served first and combining priority for high-priority queues with Weighted Fair Queuing (WFQ) for remaining queues.

Congestion Avoidance & Management preventing buffer bloat and queue exhaustion using Tail Drop or Weighted Random Early Detection (WRED) with color-aware drop probabilities and dynamic ECN (Explicit Congestion Notification).

Traffic Shaping & VLAN Queuing: Egress port shaping and per-VLAN service queuing (VLAN shaping) to prevent downstream drops.

For detailed information on OcNOS Layer 3 topics and configurations, refer to the OcNOS SP Quality of Service Guide.