Figures 5-3 and 5-4 report that on this Oracle system, over a 120-second interval, the Oracle instance consumed 10.01 seconds of CPU. This is what Oracle consumed, what Oracle's CPU requirements were, and if the same workload continues, what Oracle will consume over the next 120-second interval.

If this database server contained a single CPU core, then over the 120-second interval, the database server's CPU subsystem could supply a maximum of 120 seconds of CPU power to operating system processes. Said another way, this is the database server CPU capacity over a 120-second interval. Since Oracle consumed 10.01 seconds and the database server's capacity is 120 seconds, Oracle consumed 8.34% (10.01/120) of the available CPU. Said another way, Oracle is utilizing around 8% of the server's CPU. If this database server contains only this Oracle instance, then it is likely the operating system CPU utilization is around 12% to 18% (add between 5% to 10% for operating system overhead). This tells us that unless there are other processes consuming CPU from the database server, it is highly unlikely there is a CPU bottleneck.

In Figure 5-4, also notice that the CPU consumption is not parse time or recursive SQL time-intensive. In fact, both take less than 1% of the total CPU consumed. As I'll discuss in detail in the following chapters, when parsing or recursive SQL time starts increasing, shared pool-related latching begins to raise its ugly head.3