Sun Fire X4150 Server

Two Sun Fire X4150 servers running the OpenSolaris 2008.05 Operating System (OS) and GlassFish application server V2 posted the best Open Source result on this industry-standard Application tier benchmark. The pair of Sun Fire X4150 servers, each equipped with one Intel Xeon CPU, delivered a world record Open Source result of 1197.10 SPECjAppServer2004 JOPS@Standard. One more Sun Fire X4150 server with a single CPU was used at the database tier to run MySQL database.

This break-through result demonstrated that customers can build a complete end-to-end solution based on Open Source software from Sun that also delivers the best published price/performance score.

Benchmark Description

SPECjAppServer2004 is a multi-tier benchmark for measuring the performance of Java Platform, Enterprise Edition (Java EE) technology-based application servers that was developed to aid potential users in understanding the capabilities of different solutions under a heavy load. It measures the volume of transactions a particular system configuration can process while maintaining specified response times for 90 percent of the transactions.

SPECjAppServer2004 is an end-to-end application benchmark which exercises all parts of the underlying infrastructure that make up the application environment, including hardware, JVM software, database software, JDBC drivers, and the system network.

Benchmark Outcome

Table 1. SPECjAppServer2004 Application Tier

Sun Fire X4150 Dell PE 2950 III HP BL 460c Total number of systems 2 2 1 Total number of processors 2 4 2 Total number of Cores 8 16 8 Total space (RU) 2 4 N/A Performance (SPECjAppServer2004 JOPS@ Standard) 1,197.10 4,794.33 2,056.27

 

• Table 1 highlights the advantages of Sun's Open Source-based solution against relevant competing results. Sun's solution used the total of 2 Sun Fire X4150 servers at the application tier and one Sun Fire X4150 server at the database tier. Each server was configured with one Intel Xeon X5460 (4 cores, 1 chip) CPU for a total of 8 CPU cores at the Application tier of the benchmark.
• When compared to the application tier of Dell solution that was based on two PowerEdge 2950 servers with double the number of Intel Xeon X5470 CPUs and double the number of cores, Sun's solution takes half the physical space.
• When compared to the application tier of the HP solution that was based on one BL 460c blade module with an equivalent number of cores, Sun's solution was running community-developed and free OpenSolaris OS.
• This result demonstrates that Sun's complete open-source solution is ready for Web 2.0 and Enterprise deployments. It allows customers to run free and open-source MySQL database and GlassFish application server software on top of the OpenSolaris OS and avoid the proprietary middleware and database software lockup.

Table 2. SPECjAppServer2004 Complete Solution (All prices in $US)

Sun Fire X4150 Dell PE 2950 III HP BL 460c Performance (SPECjAppServer2004 JOPS@ Standard) 1,197.10 4,794.33 2,056.27 Total Acquisition Price $16,110 $856,725 $338,423.74 Software Acquisition Price (includes database) $0 $787,988 $280,238 External Storage Acquisition Price $0 $24,053 $28,919.74 Price/Performance (Smaller is Better) $13.46 $164.58 $178.70

 

• Table2 reflects the typical pricing structure of competitive submissions based upon proprietary databases where the majority of acquisition costs are associated with database licensing fees. Based on all current SPECjAppServer2004 benchmark submissions, Sun has the only completely Open Source-based and free software offering.
• The combination of OpenSolaris 2008.05 OS, GlassFish application server V2 and MySQL database coupled with compact Sun Fire X4150 servers demonstrates the best price/performance ever recorded on SPECjAppServer2004 benchmark.
• Additionally, the total cost of acquisition for Sun's solution is the lowest among all published SPECjAppServer2004 benchmark results. Please see the detailed pricing information here.
• Specifically, Sun's solution has 53 times better cost of acquisition than Dell's solution listed in Table 2. This results in 12 times better price/performance for Sun. In other words, while Sun's solution has about ¼ of the performance of Dell's offering, it costs less than 2% to acquire.
• Moreover, Sun's configuration has 13 times better price/performance and 21 times better cost of acquisition than HP Proliant BL460c G1 solution. In other words, while Sun's solution has about ½ of the performance of HP's offering, it has an initial cost of less than 5% of that configuration.
• While support prices are not included in Table 2, the bulk of support costs also come from database software. Notably, Sun does include MySQL Enterprise(TM) Gold subscription with Sun Fire(TM) x64 servers.
• On the database tier of the benchmark the Sun Fire X4150 server running the OpenSolaris OS and MySQL database achieved 298 SPECjAppServer2004 JOPS@Standard/core which is in the top 5 results for database per core performance.
• Notably, Sun's solution was run entirely on the internal disk drives of the server while the competition had to resort to the expensive SAN storage solutions or multi-disk JBOD arrays.
• These results confirm Sun's solution, based on Sun Fire X4150 servers, as the most price efficient platform on the planet for deployment of Java application and OLTP database services.

• View the Press Release

• View Benchmark Results

The Sun Fire X4150 server, equipped with two Quad-core Intel Xeon X5460 processors, obtained a World Record score of 303297 SPECjbb2005 bops (75824 SPECjbb2005 bops/JVM) for all dual-socket systems. Boosted by the latest version 1.6.0_05-p of Sun's Java Platform, Standard Edition (J2SE) software, this robust result was obtained under the Solaris 10 Operating System and signifies the optimization work that has been ongoing since Sun and Intel entered the landmark agreement just over a year ago.

By embracing Solaris as an enterprise class, mission critical UNIX OS for Intel Xeon processor-based systems and committing resources to make the leading performance possible across the board, Sun and Intel were able to deliver top-notch dual-socket performance on SPECjbb2005 benchmark that emulates the design of real-world server-side Java applications and provides an accurate reflection of the business logic and objects, while stressing the implementation of the Java Virtual Machine (JVM) and the scalability of the system's processors and memory. Specifically, the partnership and active collaboration between performance teams at two companies resulted in an astounding 57% gain in Java performance when compared to the result obtained with Java SE 6 (11/2006) using identical hardware configuration.

• Benchmark Results

The Sun Fire X4150 server, equipped with two Quad-core Intel Xeon X5460 processors, obtained a single JVM x86 World Record score of 277585 SPECjbb2005 bops (277585 SPECjbb2005 bops/JVM).

The SPECjbb2005 benchmark, that emulates the design of real-world server-side Java applications and provides an accurate reflection of the business logic and objects, while stressing the implementation of the Java Virtual Machine (JVM) and the scalability of the system's processors and memory, actually produces two equally important metrics - total system throughput (SPECjbb2005 bops) and JVM scalability (SPECjbb2005 bops/JVM), reflecting how different organizations may deploy their applications. Organizations and users that prefer to run a single instance of Sun's Java Platform, Standard Edition (J2SE) software, will find that the latest version (1.6.0_05-p) of the software in combination with the Sun Fire X4150 server, performs almost 2x better than the next closest score of 140220 SPECjbb2005 bops (140220 SPECjbb2005 bops/JVM) that was posted by Dell PowerEdge R200 server.

By closely collaborating on several levels, Sun and Intel were able to deliver top-shelf single-JVM performance, which was obtained under the Solaris 10 Operating System, and reflects the optimization work that has been ongoing since Sun and Intel entered the landmark agreement just over a year ago.

• Benchmark Results