AWK is what you might call a pattern processing language. It splits its input into records (delimited by newlines by default), compares each record to a number of patterns and takes appropriate actions if it matches. Each record is further split into fields (delimited by white space by default), which are assigned to the variables $1 to $NF (NF is the number of fields).
In our example, we use two patterns, namely "NR>7" and "END". NR is the number of records, a pre-defined variable. BEGIN and END are special patterns that match the beginning and the end of the input. So, what the small AWK program does is this: If the record number is larger than 7, increment the variable sum by the value of field 9 (the percentage of CPU cycles used). After all records have been processed, print out the value of sum.