MAY 2013 A PRINCIPLED TECHNOLOGIES TEST REPORT Commissioned by Red Hat, Inc. COMPARING NETWORK PERFORMANCE: RED HAT ENTERPRISE LINUX 6 VS. MICROSOFT WINDOWS SERVER 2012 For enterprises today, the ability to quickly and reliably transmit data can provide a competitive advantage. Machines, users, and applications constantly need to exchange information with one another, making the network a critical resource in a modern datacenter. Because applications do not typically manage networking resources directly and instead rely on operating systems to do so, the operating system you select may have a direct impact on TCP and UDP performance available to your applications and users. To help you quantify the actual impact, we used the Netperf benchmark in the Principled Technologies labs to compare the TCP and UDP streaming network performance of Red Hat Enterprise Linux 6 and Microsoft Windows Server 2012. We found that Red Hat Enterprise Linux 6 delivered better TCP and UDP network performance than Windows Server 2012 in most of our tests, in both out-of- box and optimized configurations. Because network performance is crucial in many business applications, selecting the appropriate operating system is critical to help you achieve your infrastructure’s maximum potential.
23
Embed
Comparing network performance: Red Hat Enterprise Linux 6 ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
MAY 2013
A PRINCIPLED TECHNOLOGIES TEST REPORT Commissioned by Red Hat, Inc.
COMPARING NETWORK PERFORMANCE: RED HAT ENTERPRISE LINUX 6 VS. MICROSOFT WINDOWS SERVER 2012
For enterprises today, the ability to quickly and reliably transmit data can
provide a competitive advantage. Machines, users, and applications constantly need to
exchange information with one another, making the network a critical resource in a
modern datacenter. Because applications do not typically manage networking resources
directly and instead rely on operating systems to do so, the operating system you select
may have a direct impact on TCP and UDP performance available to your applications
and users.
To help you quantify the actual impact, we used the Netperf benchmark in the
Principled Technologies labs to compare the TCP and UDP streaming network
performance of Red Hat Enterprise Linux 6 and Microsoft Windows Server 2012.
We found that Red Hat Enterprise Linux 6 delivered better TCP and UDP
network performance than Windows Server 2012 in most of our tests, in both out-of-
box and optimized configurations. Because network performance is crucial in many
business applications, selecting the appropriate operating system is critical to help you
5. Increase the size of the network queues from 100 to250,000 by adding the line
net.core.netdev_max_backlog=250000 to /etc/sysctl.conf and executing the command
sysctl -p
6. Stop unneeded services by running the script DisableSomeDefaultServices.sh (see the Optimizing
operating system configurations section below).
7. Reboot the server, and enter BIOS configuration.
8. Modify the BIOS settings to disable hyperthreading, and set the system power configuration to Performance per
Watt Optimized (OS).
9. Exit BIOS configuration, and boot the server.
Installing Microsoft Windows Server 2012 Datacenter 1. Insert and boot from the Windows Server 2012 Datacenter installation DVD.
2. At the first Window Setup screen, keep the defaults for installation language, time/currency format, and
keyboard input method. Click Next.
3. At the second Windows Setup screen, click Install now.
4. At the third Windows Setup screen, enter the Windows activation key, and click Next.
5. At the fourth Windows Setup screen, select the Windows Server 2012 Datacenter (Server with a GUI), and click
Next.
6. At the fifth Windows Setup screen, select the checkbox to accept the license term, and click Next.
7. At the sixth Windows Setup screen, click Custom: Install Windows only (advanced).
8. At the seventh Windows Setup screen, select Drive 2 as the Windows installation drive, and click Next to start
installation.
9. The system will reboot. At the Settings screen, enter the password for the Administrator (twice), and click Finish.
Optimizing operating system configurations Before running tests for the optimized Red Hat Enterprise Linux 6 configuration, run the following bash scripts.
Red Hat Enterprise Linux 6.4 uses the tuned utility with the throughput-performance and latency-performance profiles
to configure the server for better performance and to run the CPUs and network devices at high performance.
First, perform the second set of steps (1 through 9) in the Installing Red Hat Enterprise Linux 6.4 section on the
server under test.
A Principled Technologies test report 13
Comparing network performance: Red Hat Enterprise Linux 6 vs. Microsoft Windows Server 2012
DisableSomeDefaultServices.sh
#!/bin/bash
## For the optimized-configuration tests, disable unneeded services
## March 2013
for i in abrt-ccpp abrt-oops abrtd acpid atd auditd autofs \
Comparing network performance: Red Hat Enterprise Linux 6 vs. Microsoft Windows Server 2012
TESTING WITH NETPERF We used Netperf version 2.4.5 from ftp://ftp.netperf.org/netperf/archive/netperf-2.4.5.tar.bz2. We compiled it
for Red Hat Enterprise Linux 6 and for Windows Server 2012 as follows.
1. On the second server, running Red Hat Enterprise 6.4, make sure the GCC compile environment is installed.
yum install gcc make
2. Extract the source code from the archive.
tar jzf netperf-2.4.5.tar.bz2
cd netperf-2.4.5
3. Configure and compile netperf for Red Hat Enterprise Linux.
./configure
make all test
4. Copy the netperf and netserver binaries on each Red Hat server.
mkdir /root/np24
cp src/netperf src/netserver /root/np24
For Windows Server 2012, we chose to compile Netperf using the Cygwin environment rather than using the
Windows Driver Kit. We installed version Cygwin 1.7.18-1 under Windows Server 2012 on the server under test, and
compiled Nerperf in a Cygwin bash shell by following steps 2 and 3 above. We copied the binaries to the folder
Netperf\np24 under the Desktop.
To run the TCP and UDP throughput tests for the Red Hat Enterprise Linux server, we ran the bash script run_netserver_server1.sh (see below)on the first server, the server under test. We ran the bash script
run_netperf_STREAM_server2.sh (see below) on the second server, the traffic generator and captured its output. We rebooted both servers between runs.
To run the TCP and UDP latency tests for the Red Hat Enterprise Linux server, we ran the bash script run_netserver_server1.sh (see below)on the first server, the server under test. We ran the bash script run_netperf_RR_server2.sh (see below) on the second server, the traffic generator and captured its
output. We rebooted both servers between runs.
To run the TCP and UDP throughput tests for the Windows Server 2012 server, we ran the bash script run_netserver_windows.bat (see below) on the first server, the server under test. We ran the bash
script run_netperf_STREAM_windows.bat (see below) on the second server, the traffic generator and captured its output. We rebooted both servers between runs.
To run the TCP and UDP latency tests for the Windows Server 2012 server, we ran the bash script run_netserver_windows.bat (see below)on the first server, the server under test. We ran the bash script run_netperf_RR_windows.bat (see below) on the second server, the traffic generator and captured its output. We rebooted both servers between runs.
Comparing network performance: Red Hat Enterprise Linux 6 vs. Microsoft Windows Server 2012
APPENDIX C – DETAILED TEST RESULTS Figures 9 and 10 present detailed round-robin trip times (latency) results for our Netperf tests.
Message size (B)
TCP round-trip latency (µs)
Out-of-box Optimized Red Hat advantage
Red Hat Enterprise
Linux 6
Microsoft Windows
Server 2012
Red Hat Enterprise
Linux 6
Microsoft Windows
Server 2012
Out-of-box configuration
Optimized configuration
1 50.46 73.46 37.95 57.79 31.31% 34.33%
2 50.70 74.75 37.98 57.32 32.17% 33.74%
4 50.33 74.67 37.99 59.23 32.60% 35.86%
8 50.36 75.33 37.99 59.38 33.15% 36.02%
16 50.54 75.96 37.97 58.99 33.46% 35.63%
32 50.44 76.71 37.94 59.40 34.25% 36.13%
64 50.68 74.33 37.96 59.27 31.82% 35.95%
128 50.12 75.08 37.97 59.49 33.24% 36.17%
256 50.30 74.97 37.96 59.18 32.91% 35.86%
512 50.48 75.98 37.96 59.64 33.56% 36.35%
1,024 50.40 76.01 37.94 59.20 33.69% 35.91%
Average 50.44 75.20 37.96 58.99 32.93% 35.65%
Figure 9: TCP round-trip latency in microseconds for the two solutions, both out-of-box and optimized. Lower latencies (or higher percentage advantages) are better.
Message size (B)
UDP round-trip latency (µs)
Out-of-box Optimized Red Hat advantage
Red Hat Enterprise
Linux 6
Microsoft Windows
Server 2012
Red Hat Enterprise
Linux 6
Microsoft Windows
Server 2012
Out-of-box configuration
Optimized configuration
1 47.93 66.92 35.79 54.76 28.38% 34.64%
2 47.79 66.62 35.85 54.71 28.26% 34.47%
4 48.06 66.35 35.81 55.14 27.57% 35.06%
8 47.96 66.94 35.82 54.47 28.35% 34.24%
16 48.07 67.26 35.83 54.99 28.53% 34.84%
32 47.93 66.92 35.81 54.23 28.38% 33.97%
64 47.98 68.32 35.79 54.99 29.77% 34.92%
128 47.87 66.70 35.84 54.54 28.23% 34.29%
256 47.80 66.82 35.80 54.60 28.46% 34.43%
512 47.99 67.16 35.83 54.90 28.54% 34.74%
1,024 47.92 66.97 35.81 54.79 28.45% 34.64%
Average 47.94 67.00 35.82 54.74 28.45% 34.56%
Figure 10: UDP round-trip latency in microseconds for the two solutions, both out-of-box and optimized. Lower latencies (or higher percentage advantages) are better.
A Principled Technologies test report 23
Comparing network performance: Red Hat Enterprise Linux 6 vs. Microsoft Windows Server 2012
ABOUT PRINCIPLED TECHNOLOGIES
Principled Technologies, Inc. 1007 Slater Road, Suite 300 Durham, NC, 27703 www.principledtechnologies.com
We provide industry-leading technology assessment and fact-based marketing services. We bring to every assignment extensive experience with and expertise in all aspects of technology testing and analysis, from researching new technologies, to developing new methodologies, to testing with existing and new tools. When the assessment is complete, we know how to present the results to a broad range of target audiences. We provide our clients with the materials they need, from market-focused data to use in their own collateral to custom sales aids, such as test reports, performance assessments, and white papers. Every document reflects the results of our trusted independent analysis. We provide customized services that focus on our clients’ individual requirements. Whether the technology involves hardware, software, Web sites, or services, we offer the experience, expertise, and tools to help our clients assess how it will fare against its competition, its performance, its market readiness, and its quality and reliability. Our founders, Mark L. Van Name and Bill Catchings, have worked together in technology assessment for over 20 years. As journalists, they published over a thousand articles on a wide array of technology subjects. They created and led the Ziff-Davis Benchmark Operation, which developed such industry-standard benchmarks as Ziff Davis Media’s Winstone and WebBench. They founded and led eTesting Labs, and after the acquisition of that company by Lionbridge Technologies were the head and CTO of VeriTest.
Principled Technologies is a registered trademark of Principled Technologies, Inc. All other product names are the trademarks of their respective owners.
Disclaimer of Warranties; Limitation of Liability: PRINCIPLED TECHNOLOGIES, INC. HAS MADE REASONABLE EFFORTS TO ENSURE THE ACCURACY AND VALIDITY OF ITS TESTING, HOWEVER, PRINCIPLED TECHNOLOGIES, INC. SPECIFICALLY DISCLAIMS ANY WARRANTY, EXPRESSED OR IMPLIED, RELATING TO THE TEST RESULTS AND ANALYSIS, THEIR ACCURACY, COMPLETENESS OR QUALITY, INCLUDING ANY IMPLIED WARRANTY OF FITNESS FOR ANY PARTICULAR PURPOSE. ALL PERSONS OR ENTITIES RELYING ON THE RESULTS OF ANY TESTING DO SO AT THEIR OWN RISK, AND AGREE THAT PRINCIPLED TECHNOLOGIES, INC., ITS EMPLOYEES AND ITS SUBCONTRACTORS SHALL HAVE NO LIABILITY WHATSOEVER FROM ANY CLAIM OF LOSS OR DAMAGE ON ACCOUNT OF ANY ALLEGED ERROR OR DEFECT IN ANY TESTING PROCEDURE OR RESULT. IN NO EVENT SHALL PRINCIPLED TECHNOLOGIES, INC. BE LIABLE FOR INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES IN CONNECTION WITH ITS TESTING, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. IN NO EVENT SHALL PRINCIPLED TECHNOLOGIES, INC.’S LIABILITY, INCLUDING FOR DIRECT DAMAGES, EXCEED THE AMOUNTS PAID IN CONNECTION WITH PRINCIPLED TECHNOLOGIES, INC.’S TESTING. CUSTOMER’S SOLE AND EXCLUSIVE REMEDIES ARE AS SET FORTH HEREIN.