This value is the percentage of the last 64-second interval that the node's CPU spent actually processing traffic data, rather than running diagnostics or handling internal node-control chatter, and it applies to the 6500-family platforms. An admin watches this as a rolling load gauge on the node's processor separate from its housekeeping overhead, since a node can look 'busy' on control tasks without that reflecting real traffic load. For example, a reading of 72 in one poll cycle followed by 15 in the next suggests a short traffic burst rather than sustained overload.