White Papers

33 Sizing and Best Practices for Online Transaction Processing Applications with Oracle 11g R2 using Dell PS Series | BP1003
Also, as you can see from the IOPS activity on the disks, complete I/O activity was handled by the SSDs and
there was no I/O activity on the 10K SAS drives. This is because the array capacity utilization was 1 TB and
all the data was able to fit within the SSDs.
6.2.2 4 TB array capacity utilization
The entire database was able to fit within the SSDs in the scenario described in section 6.2.1. This is not
always the case, and the data will be spread across both SSDs and 10K SAS drives when more capacity is
used on the array. Obviously, the performance numbers will be different in this case because the I/O activity
is randomly spread across both SSDs and 10K SAS drives.
This test evaluated the scenario in which the database size was increased from 900 GB to 1.5 TB, and it
included five additional 500 GB volumes hosting other data resulting in a capacity utilization of almost 4 TB.
As a result, the data was spread across both the SSD and 10K SAS drives. The SAN HQ chart in Figure 14
shows the peak IOPS along with read and write latencies.