Network performance testing

Less than 6,000,000 PPS

1. Connect to the test and auxiliary test instances.

   For more information, see Log on to a Linux instance using Workbench.

2. On both the test and auxiliary test instances, run the following command to download Netperf.

   wget https://benchmark-packages.oss-cn-qingdao.aliyuncs.com/netperf-2.7.0.tar.gz

3. On both the test and auxiliary test instances, install Netperf and the sar monitoring tool.

   1. Run the following command to install dependencies and decompress the Netperf package.

      sudo yum install -y gcc autoconf automake libtool sysstat
      tar -zxvf netperf-2.7.0.tar.gz

   2. Run the following command to check the gcc version.

      gcc -v 2>&1

   3. If the gcc version on the test or auxiliary test instance is later than 10, complete the following steps to ensure gcc compiles correctly. Otherwise, skip this step.

      1. Run the following command to modify the nettest_omni.c file.

         cd netperf
         vim src/nettest_omni.c

      2. Press i to enter Insert mode and manually delete the declared variables from the nettest_omni.c file.

         Delete the following variable declarations:

         /* different options for the sockets        */
         
         int
           loc_nodelay,    /* don't/do use NODELAY  locally    */
           rem_nodelay,    /* don't/do use NODELAY remotely  */
           loc_sndavoid,    /* avoid send copies locally    */
           loc_rcvavoid,    /* avoid recv copies locally    */
           rem_sndavoid,    /* avoid send copies remotely    */
           rem_rcvavoid;   /* avoid recv_copies remotely    */

      3. To save the changes and exit, press Esc, enter :wq, and then press Enter.

   4. Run the following commands to compile and install Netperf.

      cd netperf
      sudo ./configure
      sudo make && sudo make install

4. On the test instance, run the following script to start 64 netserver services.

   #!/bin/bash
   for j in `seq 64`; do
       netserver -p $[16000+j] > server_$[16000+j].netperf > /dev/null 2>&1  &
   done

5. On the test instance, run the following command to query its private IP address.

   ifconfig || ip addr

   

6. On the auxiliary test instance, run the following script to send traffic to the test instance.

   #!/bin/bash
   server_ip=<test_instance_private_ip_address>
   for j in `seq 64`; do
       port=$[16000+j]
       netperf -H ${server_ip} -l ${run_time:-300} -t UDP_STREAM  -p $port  -- -m 1 -D > /dev/null 2>&1 &
   done

   Replace <test_instance_private_ip_address> with the private IP address of the test instance that you obtained in the previous step.

   

7. On the test instance, run the following command to test the packet forwarding rate.

   sar -n DEV 1

   In the test result, check the value in the rxpck/s column. The rxpck/s value indicates the total number of packets received by the test instance per second. As shown in the following figure, the average number of packets received by the test instance per second is approximately 940,000.

   

6,000,000 to 20,000,000 PPS

1. Connect to the test and auxiliary test instances.

   For more information, see Log on to a Linux instance using Workbench.

2. On the test instance and the three auxiliary test instances, run the following commands to install sockperf.

   sudo yum install -y autoconf automake libtool gcc-c++
   cd /opt
   sudo wget https://github.com/Mellanox/sockperf/archive/refs/tags/3.8.tar.gz
   sudo tar -zxf 3.8.tar.gz
   cd sockperf-3.8/
   sudo ./autogen.sh
   sudo ./configure
   sudo make -j `cat /proc/cpuinfo| grep process | wc -l`
   sudo make install

3. On the test instance, run the following command to query its private IP address.

   ifconfig || ip addr

   

4. On each of the three auxiliary test instances, run the following script to send traffic to the test instance.

   server_ip="<test_instance_private_ip_address>"
   threads=64
   msg_size=14
   run_time=60
   
   basePort=6666
   for((i=0;i<$threads;++i));do
       nohup sockperf tp -i $server_ip --pps max -m ${msg_size} -t ${run_time} --port $[${basePort}+${i}] 2>&1 &
   done

   In the script, <test_instance_private_ip_address> is the private IP address of the test instance, and run_time is the test duration. Modify these values as needed.

5. On the test instance and the three auxiliary test instances, run the following command to test the packet forwarding rate.

   sar -n DEV 1

   On the test instance, check the value in the rxpck/s column. The rxpck/s value indicates the number of packets received by the test instance per second. For example, the test instance receives approximately 12 million packets per second.

Greater than 20,000,000 PPS

1. Connect to the test and auxiliary test instances.

   For more information, see Log on to a Linux instance using Workbench.

2. On the test instance and the three auxiliary test instances, run the following commands to install sockperf.

   sudo yum install -y autoconf automake libtool gcc-c++
   cd /opt
   sudo wget https://github.com/Mellanox/sockperf/archive/refs/tags/3.8.tar.gz
   sudo tar -zxf 3.8.tar.gz
   cd sockperf-3.8/
   sudo ./autogen.sh
   sudo ./configure
   sudo make -j `cat /proc/cpuinfo| grep process | wc -l`
   sudo make install

3. On the test instance, run the following script to bind interrupt requests (IRQs).

   In high-PPS scenarios, binding interrupt requests (IRQs) to specific CPU cores keeps IRQ processing on those cores. This reduces context switching and improves processing efficiency.

   a=$(cat /proc/interrupts | grep virtio2-input | awk -F ':' '{print $1}')
   cpu=0
   for irq in $a; do
       echo $cpu >/proc/irq/$irq/smp_affinity_list
       let cpu+=2
   done

4. On the test instance, run the following command to query its private IP address.

   ifconfig || ip addr

   

5. On each of the three auxiliary test instances, run the following script to send traffic to the test instance.

   server_ip="<test_instance_private_ip_address>"
   threads=64
   msg_size=14
   run_time=60
   
   basePort=6666
   for((i=0;i<$threads;++i));do
       nohup sockperf tp -i $server_ip --pps max -m ${msg_size} -t ${run_time} --port $[${basePort}+${i}] 2>&1 &
   done

   In the script, <test_instance_private_ip_address> is the private IP address of the test instance, and run_time is the test duration. Modify these values as needed.

6. On the test instance and the three auxiliary test instances, run the following command to test the packet forwarding rate.

   sar -n DEV 1

   On the test instance, check the value in the rxpck/s column. The rxpck/s value indicates the number of packets received by the test instance per second. In the example shown in the following figure, the number of packets received per second is approximately 20 million.

Linux instances

1. Connect to the test and auxiliary test instances.

   For more information, see Log on to a Linux instance using Workbench.

2. On both the test and auxiliary test instances, run the following command to download Netperf.

   wget https://benchmark-packages.oss-cn-qingdao.aliyuncs.com/netperf-2.7.0.tar.gz

3. On both the test and auxiliary test instances, install Netperf and the sar monitoring tool.

   1. Run the following command to install dependencies and decompress the Netperf package.

      sudo yum install -y gcc autoconf automake libtool sysstat
      tar -zxvf netperf-2.7.0.tar.gz

   2. Run the following command to check the gcc version.

      gcc -v 2>&1

   3. If the gcc version on the test or auxiliary test instance is later than 10, complete the following steps to ensure gcc compiles correctly. Otherwise, skip this step.

      1. Run the following command to modify the nettest_omni.c file.

         cd netperf
         vim src/nettest_omni.c

      2. Press i to enter Insert mode and manually delete the declared variables from the nettest_omni.c file.

         Delete the following variable declarations:

         /* different options for the sockets        */
         
         int
           loc_nodelay,    /* don't/do use NODELAY  locally    */
           rem_nodelay,    /* don't/do use NODELAY remotely  */
           loc_sndavoid,    /* avoid send copies locally    */
           loc_rcvavoid,    /* avoid recv copies locally    */
           rem_sndavoid,    /* avoid send copies remotely    */
           rem_rcvavoid;   /* avoid recv_copies remotely    */

      3. To save the changes and exit, press Esc, enter :wq, and then press Enter.

   4. Run the following commands to compile and install Netperf.

      cd netperf
      sudo ./configure
      sudo make && sudo make install

4. On the test instance, run the following script to start 64 netserver services.

   #!/bin/bash
   for j in `seq 64`; do
       netserver -p $[16000+j] > server_$[16000+j].netperf 2>&1 &
   done

5. On the test instance, run the following command to query its private IP address.

   ifconfig || ip addr

   

6. On the auxiliary test instance, run the following script to send traffic to the test instance.

   #!/bin/bash
   server_ip=<test_instance_private_ip_address>
   for j in `seq 64`; do
       port=$[16000+j]
       netperf -H ${server_ip} -l ${run_time:-300} -t TCP_STREAM -p $port -- -D > /dev/null 2>&1 & 
   done

   Replace <test_instance_private_ip_address> with the private IP address of the test instance that you obtained in the previous step.

   

7. On the test instance, run the following command to test the network bandwidth.

   sar -n DEV 1

   In the test result, the rxkB/s column shows the data received in KB/s. To calculate the bandwidth in Kbps, multiply this value by 8 (since 1 byte = 8 bits).

   An example is shown in the following figure.

Windows instances

Linux instances

1. Connect to the test and auxiliary test instances.

   For more information, see Log on to a Linux instance using Workbench.

2. On both the test and auxiliary test instances, run the following commands to install sockperf.

   sudo yum install -y autoconf automake libtool gcc-c++
   cd /opt
   sudo wget https://github.com/Mellanox/sockperf/archive/refs/tags/3.8.tar.gz
   sudo tar -zxf 3.8.tar.gz
   cd sockperf-3.8/
   sudo ./autogen.sh
   sudo ./configure
   sudo make -j `cat /proc/cpuinfo| grep process | wc -l`
   sudo make install

3. On the test instance, run the following command to query its private IP address.

   ifconfig || ip addr

   

4. On the test instance, run the following command to start the service.

   sockperf sr --tcp --daemonize

5. On the auxiliary test instance, run the following command to send traffic to the test instance.

   sockperf pp -i <test_instance_private_ip_address> --tcp -t 30 -m 14 --full-log=result.json

   Replace <test_instance_private_ip_address> with the private IP address of the test instance that you obtained in the previous step.

6. On the test instance, view the tested network latency.

   An example of the test results is shown in the following figure:

   • The result line that starts with avg-latency indicates the average latency in microseconds (μs).

   • The result line that starts with percentile 99.000 indicates the 99th-percentile latency in microseconds (μs).

   

Windows instances