Oracle's SPARC T3-2 server running Oracle iPlanet Web Server middleware delivered a world record SPECweb2005 benchmark result of 113,857. Oracle's 2-socket SPARC is 9% faster than the fastest 2-socket x86-based competitive server and even 8% faster than the 4-socket HP x86-based server.
The SPARC T3-2 server with dual 1.65 GHz SPARC T3 processors using Oracle iPlanet Web Server 7.0.9 middleware delivered a world record result of 113857 on the SPECweb2005 benchmark.
This result demonstrates that the SPARC T3-2 running Oracle Solaris and Oracle iPlanet Web Server can support thousands of concurrent web server sessions and is an industry leader in web serving with a high performance and enterprise quality solution.
Oracle is the only SPECweb2005 benchmark sponsor who can demonstrate top performance using a commercially viable and production quality web serving solution with the Oracle iPlanet Web Server and the Oracle Solaris 10 operating system.
On the SPECweb2005 benchmark, the SPARC T3-2 server with two 1.65 GHz SPARC T3 processors is 8% faster than the latest Hewlett-Packard result that was just published on the HP ProLiant DL585 G7 with four 2.0 GHz AMD 6128HE processors.
On the SPECweb2005 benchmark, the SPARC T3-2 server with two 1.65 GHz SPARC T3 processors is 9% faster than the Fujitsu PRIMERGY TX300 S6 with two 3.3 GHz Intel X5680 processors.
On the SPECweb2005 benchmark, the SPARC T3-2 server with two 1.65 GHz SPARC T3 processors is 37% faster than the HP ProLiant DL370 G6 with two 3.2 GHz Intel W5580 processors.
On the Support workload of SPECweb2005, the SPARC T3-2 server with two 1.65 GHz SPARC T3 processors obtained a 41% higher score than the Fujitsu PRIMERGY TX300 S6 with two 3.3 GHz Intel X5680 processors.
The SPARC T3-2 server obtained 14.4 times the result of the 4-core IBM System p5 550 1.9 GHz POWER5+ system on the SPECweb2005 benchmark. There are no IBM POWER7 or POWER6 based system results published on the SPECweb2005 benchmark.
SPECweb2005 select results as of 8 December 2010. See the SPEC website for more. Information ordered by Result, bigger is better.
|Server||Processor||OS||SPECweb2005 Performance (\*)||Web Server|
|SPARC T3-2||2 x 1.65 T3||Solaris||113857||165024||160056||123840||iPlanet|
|HP DL585 G7||4 x 2.0 6128HE||RedHat Linux||105586||168192||175104||88576||Rock|
|Fujitsu TX300 S6||2 x 3.33 X5680||RedHat Linux||104422||162000||177000||88000||Rock|
|Sun T5440||4 x 1.6 T2 Plus||Solaris||100209||176500||133000||95000||Sun|
|Fujitsu TX300 S5||2 x 2.93 X5570||RedHat Linux||83198||106000||140000||86000||Rock|
|HP ML370 G6||2 x 3.2 W5580||RedHat Linux||83073||117120||142080||76352||Rock|
|HP DL370 G6||2 x 3.2 W5580||RedHat Linux||83073||117120||142080||76352||Rock|
|HP DL585 G5||4 x 3.1 Opt8393||RedHat Linux||71629||117504||123072||56320||Rock|
|IBM p5 550||2 x 1.9 POWER5+||SuSE Linux||7881||12240||11820||7500||Zeus|
(\*) Metrics are
Result - SPECweb2005, overall metric
Bank - SPECweb2005_banking, Banking component metric
Ecom - SPECweb2005_ecommerce, ECommerce component metric
Supp - SPECweb2005_support, Support component metric
SPECweb2005, successor to SPECweb99 and SPECweb99_SSL, is an industry standard benchmark for evaluating Web Server performance developed by SPEC. The benchmark simulates multiple user sessions accessing a Web Server and generating static and dynamic HTTP requests. The major features of SPECweb2005 are:
SPEC requires the server under test to support SSL Protocol V3 (SSLv3).
Of the various ciphers supported in SSLv3, cipher SSL_RSA_WITH_RC4_128_MD5 is currently required for all workload components that use SSL. It was selected as one of the most commonly used SSLv3 ciphers and allows results to be directly compared to each other. SSL_RSA_WITH_RC4_128_MD5 consists of:
A compliant result must use the cipher suite listed above, and must employ the 1024 bit key for RSA public key encryption, 128-bit key for RC4 bulk data encryption, and have a 128-bit output for the Message Authentication code.
All Banking workload request to the server under test use SSL, where the Ecommerce workload requests are a mix of SSL and non-SSL. Non of the Support workload requests to server under test use SSL.
When multiple 10 GbE Dual Port NICs are used, it is best practice to equally divide these NICs between PCI root nodes that are available.
Two web server instances was used. One web server instance was bound to a processor set with CPUs in the first processor chip. The other web server instance was bound to a processor set with CPUs in the second processor chip. The web server instance bound to CPUs in first processor chip was listening on the NIC IP addresses on that processor's chip PCI root node. The same was done with web server instance bound to CPUs in second processor chip. This was done to improve the locality of the processing.
Each web server is executed in the FX scheduling class to improve performance by reducing the frequency of context switches.
SPEC and the benchmark name SPECweb are registered trademarks of Standard Performance Evaluation Corporation. Results are from www.spec.org as of December 8, 2010 and this report. Oracle, SPARC T3-2, 113,857 SPECweb2005. HP ProLiant DL585 G7, 105,586 SPECweb2005. Fujitsu PRIMERGY TX300 S6, 104,422 SPECweb2005. Sun SPARC Enterprise T5440, 100,209 SPECweb2005. Fujitsu PRIMERGY TX300 S5, 83,198 SPECweb2005. HP ProLiant ML370 G6, 83,073 SPECweb2005. HP ProLiant DL370 G6, 83,073 SPECweb2005. HP ProLiant DL585 G5, 71,629 SPECweb2005. IBM System p5 550, 7,881 SPECweb2005.