Max-Burst-Size and Egress Rate Limiting

[Stephen_Elliot1]

Hello Everybody,

Can anyone explain the 'max-burst-size' variable as it relates to egress rate limiting in the command below?

configure ports rate-limit egress [no-limit | [Kbps| Mbps | Gbps] {max-burst-size [Kb | Mb]}]

More specifically, if I've been provided a Telco Ethernet WAN link with a CIR of 100Mbps over a 1Gbps bearer, what max-burst-size should I use and why?

Many thanks!
Stephen

[Hernandez__Josh]

Hello Stephen,

Great question! Max-burst-size specifies the number of consecutive bits that are allowed to be in-profile at wire-speed. This size can be specified in Kb, Mb, or Gb and will round down to the nearest supported size. This is used when rate-limiting traffic is necessary egress on a port. It maintains that should the port experience any burst of traffic at one time, it will not exceed the specified rate limit.

If this is not configured the maximum size allowed is programmed into hardware. It is recommended that the max-burst-size be configured much smaller than the expected rate of traffic.

For example, if you configure a rate limit egress on this port at 100Mbps but max-burst.size is not configured a lower rate or at all, the achieved rate will be at the maximum allowed, which on a 1Gps link could be considerably higher than the Telco company may expect or allow.

The role it plays it to keep traffic burst from exceeding the anticipated rate. See also our GTAC article:

EXOS traffic above rate-limit