Showing posts with label Performance. Show all posts
Showing posts with label Performance. Show all posts

29 June 2020

Performance Impact of ZFS Encryption on Oracle Solaris

Transparent Encryption is very easy to use on Oracle Solaris.
You just need to set the encryption property when you create a new filesystem and provide a passphrase or keyfile.

On a SPARC S7 LDom we have 3 ZFS filesystems with different encryption settings.

# zfs get encryption v0123_db/plain v0123_db/encr v0123_db/encr256
NAME              PROPERTY    VALUE        SOURCE
v0123_db/encr     encryption  on           local
v0123_db/encr256  encryption  aes-256-ccm  local
v0123_db/plain    encryption  off          -

Now lets see how much is the difference in write performance if we copy a 1 GB file.

# ls -lh p25604852_1100_Solaris86-64_1of4.zip
-rw-r--r--   1 marcel   staff       1.3G Apr  7  2017 p25604852_1100_Solaris86-64_1of4.zip
#

# time cp p25604852_1100_Solaris86-64_1of4.zip /plain

real    0m8.829s
user    0m0.002s
sys     0m1.711s

# time cp p25604852_1100_Solaris86-64_1of4.zip /encr

real    0m9.229s
user    0m0.002s
sys     0m1.747s

# time cp p25604852_1100_Solaris86-64_1of4.zip /encr256

real    0m9.733s
user    0m0.002s
sys     0m1.754s

The difference is a low one digit percent value.

Performance impact is a little larger when doing a simple read test.

# time cp /plain/p25604852_1100_Solaris86-64_1of4.zip /tmp

real    0m4.216s
user    0m0.002s
sys     0m3.810s

# time cp /encr/p25604852_1100_Solaris86-64_1of4.zip /tmp

real    0m5.131s
user    0m0.003s
sys     0m5.028s

# time cp /encr256/p25604852_1100_Solaris86-64_1of4.zip /tmp

real    0m5.400s
user    0m0.003s
sys     0m5.287s

Learn more about ZFS encryption with the Oracle Solaris 11.4 ZFS Admin Guide

https://docs.oracle.com/cd/E37838_01/html/E61017/gkkih.html

21 June 2017

SPARC M12 and S7 CPU comparison using SLOB



Fujitsu SPARC M12 Server

Fujitsu recently announced new SPARC M12 Servers using the SPARC64-XII CPU. This systems hold several performance world records. Check out details at http://www.fujitsu.com/global/products/computing/servers/unix/sparc/key-reports/benchmarks/

But how does this CPU and system scale and how is the performance compared to our own Oracle SPARC S7 Server?

SLOB DB Benchmark

To compare the two systems I setup a SLOB (Silly Little Oracle DB Benchmark) environment.

The SLOB benchmark executes 500'000 SQL select statements (SLOB Ops). The SGA is 20GB is size to make sure all data is in the Database Cache and no physical I/O is required. This way we measure CPU, Memory, OS and DB Performance.


SLOB results

The 12-core SPARC64-XII scales very well. Using 96 parallel readers we reach 78'000 SLOB OPS per second per socket. This is 2x the SLOB OPS compared to the 8-core SPARC-S7 CPU. 

Calculating the performance down to 1 core, we see a peak of 6500 SLOB OPS per second per core on the M12 and 5000 SLOB OPS per second per core on the S7. A M12 core outperforms the S7 core by 30%. On the S7 we see better results if only 1 single reader is executed.


Technical details

To make sure we compare oranges with oranges, the same setup was used on both servers.
A Logical Domain was created using 48GB RAM and 1 socket assigned to it.
Solaris 11.3 SRU19 / Oracle DB 12c DATABASE BUNDLE PATCH: 12.1.0.2.170418 (25397136)

We used a SPARC S7-2 (4.267 Ghz) and a SPARC M12-2S (4.25 Ghz).

08 September 2016

Performance Comparison SPARC T4 and SPARC S7

SPARC S7
JomaSoft replaces the SPARC T4-1 Server with the new SPARC S7-2 Server (with 2 sockets / 8 cores each at 4.26Ghz).

Read more about the SPARC S7-2 Server

Comparison
We created a 3GB / 1 core LDom on our SPARC T4-1 Server running Solaris 11.3 SRU11 with all data stored on a SAN Disk. Inside the LDom we installed our VDCF application and loaded datacenter configuration into the VDCF sqlite database. Next we executed datacenter analysis like patch comparison, calculated migration possibilities and server configuration consistency checks. The analysis are traditional single thread workload.

After the tests on the SPARC T4-1 we migrated our LDom to the new SPARC S7-2 Server. This allowed us to compare the systems using the same Operating System, Setup and Data, to make sure we compare “apple to apple”.

The results for our workload showed a 2x faster performance on the SPARC S7-2 Server. We are very happy with this results. This workload did not use the Software In Silicon features, only performed better because of the new CPU architecture (higher frequency, more and better CPU cache and memory).

In my view the SPARC S7-2 Server is the ideal platform for customers to replace their old SPARC hardware with an excellent price / performance ratio.