Congestion Management; Overview; Congestion Management Policy - H3C S5120-EI Series Operation Manual

5

Congestion Management

When configuring congestion management, go to these section for information that you are interested
in:

Overview

Congestion Management Policy

Configuring an SP Queue
Configuring a WRR Queue
Configuring a WFQ Queue
Configuring SP+WRR Queues
Displaying and Maintaining Congestion Management
Overview
When the rate at which the packets arrive is higher than the rate at which the packets are transmitted on
an interface, congestion occurs on this interface. If there is not enough storage space to store these
packets, parts of them will be lost. Packet loss may cause the transmitting device to retransmit the
packets because the lost packets time out, which causes a malicious cycle.
The core of congestion management is how to schedule the resources and determine the sequence of
forwarding packets when congestion occurs. Congestion management processing includes queue
creating, traffic classification, packet enqueuing, and queue scheduling.
Congestion Management Policy
Queuing technology is generally adopted to solve the congestion problem. The queuing technology is to
classify the traffic according to a specified queue-scheduling algorithm and then use the specified
priority algorithm to forward the traffic. Each queuing algorithm is used to solve specific network traffic
problems and affects the parameters such as bandwidth allocation, delay and delay jitter.
The S5120-EI series support the following four queue scheduling methods:
Scheduling all queues with the strict priority (SP) algorithm.
Scheduling all queues with the weighted round robin (WRR) algorithm.
Scheduling all queues with the weighted fair queuing (WFQ) algorithm
Scheduling some queues with the SP algorithm and some with the WRR algorithm.
This section describe how SP, WRR, WFQ, and SP+WRR work in details.
1) SP queue-scheduling algorithm
5-1