LANforge-FIRE Cookbook

The LANforge-FIRE Cookbook provides a set of high-level examples of how to setup useful test scenarios in LANforge-FIRE for traffic generation. Each example intends to give the reader a brief introduction to the test scenario and a set of step-by-step instructions on how to use the LANforge-GUI to configure the test.

All of the following examples will work on Linux systems running the LANforge software with the LANforge kernel and a sufficient license. If you are running another Linux kernel, you will not be able to exactly duplicate some of the examples, but there are usually work-arounds available to assist you. Please contact us at support@candelatech.com if you have any questions.

If you are using the Windows version of LANforge, you will have to modify ports using the Windows utilities. Everything else should work approximately the same, but the performance is limited to 10Mbps speeds.

LANforge-FIRE Traffic Generation

Before attempting the examples below, ensure that you have successfully followed these software installation guides:

It is also recommended that you back up your current running LANforge Server database so that you may safely return to your current operating state.

LANforge-FIRE Cookbook Examples

1. Generating Traffic to a Switched Network

Goal: Setup and run traffic on a switched network. Return to LANforge-FIRE Cookbook

In this test scenario, a basic, unmanaged switch is the device under test (DUT) while LANforge-FIRE is used to generate traffic and verify switch throughput.

Connect two available data generating ports of your LANforge Server to the DUT, an unmanaged switch.

Setup the LANforge Ports so that they have valid IP addresses.

A: Go to the Port Manager

B: Modify port for Endpoint A (eth2)

C: Modify port for Endpoint B (eth3)

D: Verify the port configuration

For more information, see the LANforge User's Guide: Ports (Interfaces)

A: Go to the Port Manager
B: Modify port for Endpoint A (eth2)
C: Modify port for Endpoint B (eth3)
D: Verify the port configuration

A: Go to the Layer-3 tab
B: Create a new Cross-Connect
C: Verify the new Cross-Connect

A: Select the cross-connect on the Layer-3 tab, click Start and then Display
B: View the Layer-3 cross-connect display

A: Select Report Manager from the Reporting pull-down menu
B: Choose a directory to save raw (.csv) data then click Start Saving. After desired test duration, click Stop Saving and Generate Report. NOTE: Your report will look cleaner if you stop traffic about 2 minutes before you stop saving data.
C: Choose a directory to save the HTML report, select the entities to be included, then click Generate Report
D: This window lists all the reports being generated. The last line indicates the location of the HTML report.
E: View the HTML report

A: Go to the Port Manager
B: Modify port for Endpoint A (eth2). Set a valid network IP address and Gateway IP.
C: Modify port for Endpoint B (eth3). Set a valid network IP address and Gateway IP.
D: Verify the port configuration

A: Go to the L3 Endps tab
B: Scroll to the right to view Latency and Dropped Packets

A: On the Port Manager tab, setup the LANforge ports with valid IP addresses.
B: On the Layer-3 tab, create a UDP connection. • Set Endpoint Side-A to use the DUT WAN port and Endpoint Side-B to use the DUT LAN port, then select CX Type LANforge UDP then select Apply.
C: Select the Script button to setup the scripting parameters. • Setup the script to iterate over the rates and payload sizes to be tested. • For more information on this step, see the LANforge FIRE Cookbook example: Scripted Layer-3 Test
D: Highlight the connection and select the Start button.
E: The final test report shows the results of the test run. Here we can see that the DUT has the best bi-directional throughput with 1460Byte payloads at 24.9Mbps. Full script report for the UDP test.

A: Highlight, then modify the previous connection, change the Name then select CX Type LANforge TCP, then Apply to create the new connection.
B: Select the Script button to setup the scripting parameters. • Setup the script to iterate over the rates and payload sizes to be tested. • For more information on this step, see the LANforge FIRE Cookbook example: Scripted Layer-3 Test
C: Highlight the connection and select the Start button.
D: The final test report shows the results of the test run. Here we can see that the DUT has the best bi-directional throughput at with 1472Byte payloads at 25.3Mbps. Full script report for the TCP test.

A: On the Port Manager tab, create 5 MAC-VLANs on the LANforge Port connected to the DUT LAN port.
B: Verify that the MAC-VLANs have correct IP addresses.
C: Create a Layer-3 connection that has a low-speed rate with Multi-Conn set to 10000 on Endpoint Side-A. Multi-Conn should be set to 1 on Endpoint Side-B. • Side-A will be one of the MAC-VLANs and Side-B will be the port connected to the DUT WAN port. This setup will initiate the TCP sessions from the LAN side of the DUT. • Low-speed depends on the DUT, we could also set the rate to zero which would allow the TCP connections to be setup without payload data to be transmitted, but this would not give an accurate picture of the firewall performance. Here we are using 1Kbps connections with 1KB size payload. • This is an iterative test, the number of TCP connections to use will depend on the DUT capabilities. Modify the number of connections as necessary to find the most accurate measurement. • The DUT should be power-cycled to reset it before each test run.
D: Select the Batch-Create button to create 4 more copies of this connection each with a new MAC-VLAN port.
E: Highlight and Start each set of 10000 connections until the target max simultaneous connections are running.
F: On the Layer-3 Endpoints tab, highlight the Running A-Side Endpoints, then right-click and select Calculations.
G: The top line, Sum, is what we are interested in for Maximum Concurrent TCP Connections.
H: Scroll right to the CX Active and CX Established columns and select the Refresh button. This DUT can maintain a maximum of 41,864 simultaneous TCP connections. • CX Active is the metric we are attempting to measure, Maximum Simultaneous TCP Connections. It will fluctuate with the DUT's ability to maintain the number of active TCP connections. • CX Established is the number of TCP connections LANforge has established since the start of the test. It will continue to increase as the DUT closes the TCP connections it cannot maintain.

A: Create a Layer-3 TCP connection with the Duration and IP Port set to zero.
B: Highlight and Start the connection.
C: View the Connections/second rate on the Layer-3 Endpoints tab. This DUT can setup about 120 Connections/second.

A: CTRL-select all of the virtual wireless stations and eth0 then click Modify • NOTE: This will bring up 5 modify windows at once
B: Verify the 'DHCP-IPv4' checkbox is selected for each interface intended as a DHCP client (if not done when created): • If using WEP for the wireless stations, replace 'none' in the Key field with the hexadecimal key string generated by your access point where you enabled WEP • Click OK when finished
C: If your access point DHCP service is running, the 4 virtual wireless stations and eth0 will acquire their IP addresses after a few seconds
D: Scroll to the right to view each station's link quality and other interface details

A: Go to the Layer-3 tab and click Create
B: Create a wireless-to-wired UDP speed test: • Test 1: sta1-eth0, UDP, 15 Mbps bi-directional, 1472 byte packets
C: Create a wireless-to-wired TCP connections per second test: • Test 2: sta2-eth0, TCP, Endpoint-A IP Port 0, CX Duration 0 • NOTE: Duration 0 means there will be zero delay between attempted TCP connections created between sta2 and eth0 which will allow LANforge to generate as many connections per second as possible through the access point
D: Create a wireless station-to-wireless station speed test: • Test 3: sta3-sta4, UDP, 1.544 Mbps bi-directional, 1472 byte packets

A: Test 1 results show that the wireless station can transmit 15 Mbps, but can only receive about 10 Mbps from the access point
B: Test 2 results show about 400 TCP connections per second are being established through the access point
C: Test 3 results show a maximum throughput of about 1.5 Mbps between wireless stations through the access point • NOTE: This test was run iteratively at different speeds to determine that 1 Mbps was the maximum stable throughput between wireless stations through the access point

A: Select the Armageddon connection then click Start
B: It will take a moment for the Armageddon traffic to stabilize and depending on the hardware, LANforge will settle on an actual rate which may differ from the requested rate
C: In this example, this server (3.2GHz Xeon with 4-port 10/100/1000 PCI-X) is capable of reaching 81,000pps with 1514byte packets or, an equivalent speed of about 981Mbps bi-directional. • NOTE: Delay for Armageddon connections is measured in microseconds and in this example, the server experiences about 2900us of delay when sending to itself.

A: Go to the Port Manager
B: Modify the port connected to the Web Server. Set a valid network IP address and Gateway IP.
C: Verify the port configuration

A: Go to the Port Manager
B: Modify the port connected to the DSLAM. Set a valid network IP address and IP mask.
C: Verify the port configuration

A: On the Port Mgr tab, select the port that will connect to the multi-port switch and will also be your VLAN trunk. This is the LF Edge part of the diagram.
B: Click the Create button: • Select the 802.1Q-VLAN virtual interface type • The VLAN ID should correspond to your multi-port switch VLAN configuration • Enter Quantity (number of virtual interfaces to be created) • Enter the appropriate IP address and IP mask • Click OK or Apply to create the virtual interfaces
C: Verify that the virtual interfaces are created and have the correct IP assignments
D: Modify each port's MAC address for each virtual interface so that they are unique • Select the 'Set MAC' checkbox on the left to edit the MAC address field
E: Scroll to the right to verify that the MAC addresses are unique

A: Select the connections you want to start
B: Click the Start button

A: Go to the Port Manager
B: Modify eth2 and eth3 to set a valid network IP address and mask • If your network has DHCP service, you can select the 'DHCP-IPv4' checkbox so that each port is a DHCP client and will acquire its IP address from your DHCP server
C: Verify the port configuration

A: Select Test-1 and click the Start button
B: Go to the VoIP/RTP Endps tab to see detailed results: • The PESQ score will be reported after the first successful call is completed and updated after each subsequent call • NOTE: Endpoints are Unregistered while the call is in progress because they are not calling through the VoIP gateway
C: Stop Test-1, select Test-2 and click Start
D: Go to the VoIP/RTP Endps tab to see detailed results: • PESQ remains 0: 0 when it is disabled for the call in progress • NOTE: Endpoints are Registered with the VoIP gateway while the call is in progress • Calls Attempted, Calls Completed and Calls Failed can be viewed by scrolling to the right on the VoIP/RTP Endps tab

A: Go to the Port Manager and select port ath0 or wmaster0
B: Modify port ath0 or wmaster0
C: Setup the Wifi Channel and Mode you want to use, then click OK

A: Go to the Port Manager and select ports eth2 and eth3
B: Modify ports eth2 and eth3: • In this example, eth2 and eth3 are physically connected with a patch cable
C: Configure each port with a valid IP address, then click OK

A: Go to the Port Manager and select ports eth2 and eth3
B: Modify ports eth2 and eth3 • In this example, eth2 and eth3 are connected to another LANforge system running a WanLink so that the Armageddon traffic can be sniffed on the other machine's interface
C: Configure each port with a valid IP address, then click OK • NOTE: Be sure that both ports are in Promiscuous mode by selecting the 'Set PROMISC' and 'PROMISC' checkboxes

A: On the Armageddon tab, click Create
B: Select ports eth2 and eth3 and enter the speed and packet size
C: Enter values for the Source and Destination MAC addresses and specify a MAC count

A: Select the Armageddon connection then click Start
B: On the Port Mgr tab of the other LANforge system, select one of the physical interfaces in the Armageddon connection
C: Click Sniff Packets to launch Wireshark and begin sniffing traffic. • Stop the Wireshark capture after a few seconds via the stop icon or pull-down menu (Capture>Stop)
D: Select several packets and note their MAC addresses
E: Verify that the MAC addresses for each packet are different

A: Highlight the Layer-3 connection and select Modify.
B: Select the Script button on Endpoint A.
C: Select the Script Type, RFC-2544. Note: A default set of payload sizes are setup based on RFC-2544 but, can be changed by typing over the default values. Note: For Layer-3 UDP and TCP connections, 'payload size' refers to size of the payload being carried by the protocol and not the ethernet frame size.

A: Select Symmetric for the script to run both endpoints for a bi-directional traffic test.
B: Set the Run and Pause Duration.
C: Note the total number of Script Iterations and Estimated Total Duration to help determine how long it will take to run this script.
D: Select OK to close the Add/Modify Script window.
E: The Script Type for Endpoint B is set to NONE because Endpoint A is controlling both ends of the connection in this symmetric script example.
F: Select OK to close the Create/Modify Cross Connect window.

A: Highlight the Layer-3 connection and select Start.
B: A script report window will pop up and show the details of each iteration of the scripted connection as it is run.
C: At the conclusion of the script, the report window will display a summary of the entire scripted connection results. Full Script Report for this example. - per iteration details - raw CSV data for all iterations - spreadsheet matrices for creating your own 3D graphs - system information

A: Highlight the Armageddon connection and select Modify.
B: Select the Script button on Endpoint A.
C: Select the Script Type, RFC-2544. Note: A default set of payload sizes are setup based on RFC-2544 but, can be changed by typing over the default values. Note: For Armageddon UDP connections, 'payload size' refers to size of the ethernet frame size.

A: Select Symmetric for the script to run both endpoints for a bi-directional traffic test.
B: Set the Run and Pause Duration, modify Rates and Payload Sizes as needed.
C: Note the total number of Script Iterations and Estimated Total Duration to help determine how long it will take to run this script.
D: Select OK to close the Add/Modify Script window.
E: The Script Type for Endpoint B is set to NONE because Endpoint A is controlling both ends of the connection in this symmetric script example.
F: Select OK to close the Create/Modify Cross Connect window.

A: Using the LANforge GUI, go to the Port Manager tab.
B: Assign IP addresses and masks to the two ports.
C: Verify the port configuration.

A: Go to the Layer-3 tab, and select the Create button.
B: For Endpoint A, type in a name in the CX Name field, choose a CX Type of LANforge/TCP, choose the correct ports and set the Report Timer to 1000 ms.
C: Set the Min Tx Rate and Max Tx Rate on both Endpoints A and B to 9.6Kbps. NOTE: Higher speeds and/or bursty transmit rates will require more memory to process per connection.
D: Leave the Min Pkt Size and Max Pkt Size on both Endpoints A and B set to 1024B. NOTE: Larger packet sizes will use more memory, smaller packet sizes will require more CPU processing time.
E: Leave the TCP Duration set to Forever, unless you want each of the multiple TCP connections to teardown and restart after a specified duration.
F: For Endpoint A, set the Multi-Conn to the desired number of multiple TCP connections. You can type in any value that your license allows here or choose one from the drop down box. The recommended approach is to create a new Multi-Conn connection for every 5000 connections.
G: For Endpoint B, set the Multi-Conn to one, then select OK to create the connection. Endpoint B is the "server" side of the TCP connection, therefore it is only necessary to have one Multi-Conn setup on Endpoint B to reply to all of the Endpoint A TCP SYN packets when the 5000 connections are being established.

A: On the Layer-3 tab, highlight the first connection created in the previous step and select Modify.
B: Select the Batch-Create button.
C: Set the Quantity to 5, and set Port Increment A and Port Increment B to 0.
D: Select the Apply button.
E: Verify that six Multi-Conn TCP connections were created.

A: Highlight all six connections.
B: Select the Start button
C: Go to the L3 Endps tab to verify that all 30,000 connections are running.
D: Or, go to the Port Mgr tab, and sniff one of the ports while the Multi-Conn TCP connections are running. NOTE:You will need a lot of memory to run a Wireshark capture on 30,000 connections. It is probably best to stop all connections, set up the Wireshark capture, start all connections and let them run for 30 seconds to 1 minute before stopping them all, then stop the Wireshark capture and wait for it to process all those packets!
E: After the Wireshark capture is complete, go to Statistics - Conversations to allow Wireshark to analyze the conversations and show that all 30,000 connections were captured. This will take some time to complete depending on the size of the capture.