By Franck Pachot

.
In the latest post I’ve run a cached SLOB workload on Oracle Cloud IaaS to measure logical reads per seconds on a system covered by 2 processor licences (so 4 OCPs). Just as a comparison, here is the same on Oracle PaaS database as a service.

PaaS

The CPUs in PaaS are not exactly the same: E5-2690 v2 (3.00GHz) – it was E5-2699 v3 (2.30GHz) for my IaaS test.


[oracle@DBI122 ~]$ lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                2
On-line CPU(s) list:   0,1
Thread(s) per core:    1
Core(s) per socket:    2
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 62
Stepping:              4
CPU MHz:               2992.874
BogoMIPS:              5985.74
Hypervisor vendor:     Xen
Virtualization type:   full
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              25600K
NUMA node0 CPU(s):     0,1
[oracle@DBI122 ~]$ cat /proc/cpuinfo | tail -26
processor       : 1
vendor_id       : GenuineIntel
cpu family      : 6
model           : 62
model name      : Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz
stepping        : 4
microcode       : 0x428
cpu MHz         : 2992.874
cache size      : 25600 KB
physical id     : 0
siblings        : 2
core id         : 1
cpu cores       : 2
apicid          : 2
initial apicid  : 2
fpu             : yes
fpu_exception   : yes
cpuid level     : 13
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx rdtscp lm constant_tsc rep_good nopl eagerfpu pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm xsaveopt fsgsbase smep erms
bogomips        : 5985.74
clflush size    : 64
cache_alignment : 64
address sizes   : 46 bits physical, 48 bits virtual
power management:

So it seems that PaaS has faster CPU (see frequency and bogomips) but nothing worth a real test:

SLOB

Here I’ve run 1 to 8 SLOB sessions as I did in the previous post and here is the result:


Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               1.0              25.1      0.00      2.19
              DB CPU(s):               1.0              25.1      0.00      2.18
  Logical read (blocks):         611,210.2      15,357,878.4
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               2.0              40.1      0.00      7.71
              DB CPU(s):               2.0              40.1      0.00      7.70
  Logical read (blocks):       1,195,863.3      24,031,350.5
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               3.0              75.1      0.00     11.86
              DB CPU(s):               3.0              75.0      0.00     11.84
  Logical read (blocks):       1,720,446.4      43,208,149.8
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               4.0              70.7      0.00     11.78
              DB CPU(s):               4.0              70.6      0.00     11.76
  Logical read (blocks):       2,266,196.4      40,174,995.7
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               5.0             125.1      0.00     13.17
              DB CPU(s):               5.0             124.9      0.00     13.15
  Logical read (blocks):       2,802,916.0      70,385,892.6
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               6.0              90.1      0.00     15.80
              DB CPU(s):               6.0              90.0      0.00     15.78
  Logical read (blocks):       3,312,050.8      49,898,529.6
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               7.0              95.5      0.00     17.22
              DB CPU(s):               7.0              95.3      0.00     17.18
  Logical read (blocks):       3,812,912.2      52,225,112.1
 
Load Profile                    Per Second   Per Transaction  Per Exec  Per Call
~~~~~~~~~~~~~~~            ---------------   --------------- --------- ---------
             DB Time(s):               8.0             141.3      0.00     16.45
              DB CPU(s):               7.9             140.2      0.00     16.33
  Logical read (blocks):       4,237,433.6      75,154,623.7

Faster CPU but lower logical reads processed by seconds… Don’t look only at the specs when choosing an instance type. Test it with your workload…

Besides performance, I really like the Oracle Cloud PaaS for Database. You have easy provisioning (a few clicks) but still full access (root, grid, oracle, sysdba). There is no competitor on that. In other clouds, either you go IaaS and you have to install and configure everything yourself, or you go PaaS and you have very limited admin access. Here you have both.

Post Views: 475
Thumbnail [60x60]
by
Oracle Team