>It's more than 20; there's 27 rate-limit statements on an interface
>we use. I think the limit is 100, and is determined by the range of
>numbers allowed for extended IP access lists (100 to 199).
Total limit of independed rate-limit -commands per interface is 100.
I have used more than 90 per interface without a problem.
There is also a limitation of cascaded rate-limits, where you
use the same acl and rate-limit it several times (to set type-of-service
or something else). Don't remember what was the exact limitation
here but I recall that it was as low as three. Question is what happens
to the flow after this 3 rate-limit "hops" is done. Is it dropped or
forwarded normally?
Example:
rate-limit input access-group 101 2048000 9188 91880 conform-action transmit exceed-action set-prec-continue 1
rate-limit input access-group 101 1024000 9188 91880 conform-action transmit exceed-action set-prec-continue 2
rate-limit input access-group 101 512000 9188 91880 conform-action transmit exceed-action set-prec-continue 3
rate-limit input access-group 101 256000 9188 91880 conform-action transmit exceed-action set-prec-continue 4
rate-limit input access-group 101 128000 9188 91880 conform-action transmit exceed-action drop
This gives to a given flow (acl 101 describes it) 2 Mbps speed with whatever TOS setting it might have,
then it maps TOS to 1 and gives 1 Mbps speed to it. Again the exceeded flow rate is mapped to TOS 2
and given 512 kbps bw.
Does anybody remember the limit of CAR shaping per acl what is allowed?
Jorma
------------------
CCIE#4185
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:12:15 EDT