Chapter 37: Service Class Names
STANDARD Revision 1.0
C4® CMTS Release 8.3 User Guide
© 2016 ARRIS Enterprises LLC. All Rights Reserved.
1003
Calculating Average Usage
Periodically the C4/c CMTS adds the counts for each service flow that is being monitored to the running total for that
sampling period. When the number of system timer cycles is sufficient to equal the configured sampling interval, the C4/c
CMTS converts the total count for those cycles to kbps and overwrites the count of the oldest sampling interval in the
history. It then calculates the average kbps over the whole history, which would be just one sampling period if the feature
has just been enabled or the modem has just registered. The history could be as long as one day, the maximum allowed by
the feature. It then compares this average usage over the length of the available history to the penalty threshold to decide
whether to trigger the Enforce Rule.
Feature Interactions
The effect of enabling this feature or of applying an Enforce Rule to a given CM may not be immediately apparent because
ISCA relies on system timers. Allow up to two minutes for configuration changes to take effect.
The stale-mac parameter of the configure aging command has a default value of 7 days, which is probably far longer
than any penalty an MSO would enforce. However, if the stale-mac parameter were given a value less than that of an
ISCA penalty duration, then the ISCA feature would not behave as expected with respect to penalized CMs that reboot.
Customers who disable ISCA on a heavily loaded C4/c CMTS should allow ten minutes for all penalties and histories to
clear.
System overload conditions may affect ISCA penalties because the CER processes ISCA more slowly than expected. The
CER does not make any penalty enforcements or removals while it is in an overload state. However, it will maintain
accurate counts and time records during that state. Once the overload condition has cleared, then at the next sampling
interval CMs can now be removed or added to the penalty list.