This section describes how to configure HQoS based on a main interface.
When congestion occurs because the assured bandwidth of traffic exceeds the interface bandwidth, you can change the default CIR, PIR, CBS, PBS, and weight of user queues on a main physical interface, thereby relieving congestion. If all trunk member interfaces are on the same network processor (NP) of the same board, the trunk member interfaces share the bandwidth configured on the trunk interface. Otherwise, the rate limit of each trunk member interface is the one configured on the trunk interface. In this case, the rate limiting fails. Therefore, users need to configure a bandwidth distribution mode so that traffic is allocated to trunk member interfaces based on weights.
In VS mode, this configuration is supported only by the admin VS.
The system view is displayed.
The interface view is displayed.
User queue scheduling parameters are configured to implement HQoS for user services.
HQoS is configured to take effect for L3VPN, IP, VLL, VPLS, and EVPN traffic on the interface.
Weight-based bandwidth allocation for a trunk's member interfaces is configured for HQoS scheduling.
The HQoS function is enabled for the traffic carried by a BGP local IFNET tunnel on the interface.
After HQoS is configured on an outbound interface of a BGP local IFNET tunnel, if the outbound interface resides on a downstream board equipped with a non-eTM subcard, you need to perform this step to allow HQoS to take effect.
The default user queue scheduling parameters are modified.
The configuration is committed.
Return to the system view.
The slot view is displayed.
The minimum default burst size and burst time are configured for user queues.
The function of automatically adjusting the number of users on a channel is disabled.
The configuration is committed.