User's Manual

Table Of Contents

Port based ingress rate limiting, see also the Functional diagram

Fig. 7.25: Menu Switch settings - PIRL

The device supports per port TCP/IP ingress rate limiting along with independent Storm prevention.

Port based ingress rate limiting accommodates information rates from 64 Kbps to 1 Mbps in increments

of 64 Kbps, from 1 Mbps to 100 Mbps in increments of 1 Mbps and from 100 Mbps to 1000 Mbps in

increments of 10 Mbps.

In addition to this, the device supports Priority based ingress rate limiting. A given ingress rate resource

can be configured to track any of the four priority traffic types. One of the popular schemes for imple-

menting rate limiting is a leaky bucket. The way a leaky bucket scheme works is that the bucket drains

tokens constantly at a rate called Committed Information Rate (CIR) and the bucket gets replenished

with tokens whenever a frame is allowed to go through the bucket. All calculations for this bucket are

done in tokens. Therefore, both bucket decrementing and incrementing is performed using tokens (i.e.,

frame bytes are converted into bucket tokens for calculation purposes).

The device supports a color blind leaky bucket scheme.

The traffic below Committed Burst Size limit (CBS Limit) is passed without any further actions. If the

traffic burst were to continue and the bucket token depth approaches closer to the Excess Burst Size

limit (EBS Limit) by less than the CBS Limit, then a set of actions are specified. Note that if the frame

gets discarded then the equivalent number of tokens for that frame will not get added to the bucket.

There are the two default ingress limiting rules already configured in the switch default configuration.

They limit the maximum allowed ARP traffic comming to the CPU port to 10Mbps from Eth1 and 10Mbps

from Eth2 ports.

Configuration