Often network ports are utilized at rates below 50%. If several network ports are aggregated, there is a risk of oversubscribing the tool ports. Alarm buffer thresholds are used to monitor the congestion within the GigaVUE node caused by microbursts or by oversubscription of tool ports.
The buffer usage on any port remains at zero until the maximum line rate of the port is reached. When the usage crosses 100% either instantaneously, in the microburst case, or prolonged, in the oversubscription case, there is congestion.
The internal buffer on the GigaVUE node can absorb a certain number of packet bursts. During congestion, packets are buffered in the chassis and the buffer usage is reported on the corresponding ports and in the corresponding direction: rx (ingress) and tx (egress).
Reporting the buffer usage provides a trend of how the microbursts are causing congestion, so more tool ports can be added before packets are dropped. Buffer usage is measured in intervals of 5 seconds. The peak buffer usage within a 5-second interval is reported. Use the show profile commands to see trends of buffer usage over time.
When buffer usage is less than or equal to zero, there is no congestion, so no packets are dropped due to buffer unavailability.
When buffer usage is greater than zero, there is congestion. When buffer usage is greater than zero on any port in any direction, there is a chance that the packets (that caused the buffer usage to increase) are dropped due to unavailable buffers. However, it is unlikely to see packet drops due to buffer unavailability when the buffer usage on a port is less than 5%.
The buffer usage feature is supported on all ports and module types on the GigaVUE‑HC3 and GigaVUE‑HC2 (equipped with Control Card version 1 only).
Refer to the following sections for configuring buffer thresholds and for configuring a notification that can be sent when a threshold is exceeded:
|Set Alarm Buffer Thresholds|
|Buffer Usage Alarm|