Using NetStress

In order to measure throughput performance between two nodes on a network we need a transmitter and receiver.  The transmits packets to the  receiver -- and the receiver, in turn, sends the result back to the transmitter.  The results represent a quantifiable metric that reflects the performance of the path between the transmitter and receiver machines. By selecting the transmitter and receiver machines at various points within the network one can analyze critical portions of the data path.

Each instance of NetStress can run as both a transmitter and receiver, or just as a receiver.  If you wish for a particular instance of NetStress to run just as a receiver then it is not necessary to specify a remote receiver IP address.  On the other hand, if you wish for a particular instance of NetStress to run as both a transmitter and receiver, then use the 'Remote Receiver IP' button along the main menu to select or enter the IP address of a remote receiver.  NetStress uses a UDP broadcast mechanism to automatically discover other instances of NetStress on the network that can act as a remote receiver.  If no remote receivers are listed below this is probably because NetStress is not running on another machine.  You can also manually enter the IP address of a remote receiver.  There is one requirement -- the remote receiver must be reachable from the transmitter using ICMP (i.e. ping).

Running NetStress:

1) Launch NetStress on machine A -- call this "Node A". A dialog appears with a list of available network interfaces -- choose one.

2) Launch NetStress on machine B -- call this "Node B". A dialog appears with a list of available network interfaces -- choose one. Be aware that in order for the two instances of NetStress to be able to communicate then the interfaces you've selected must be on the same network. If there is only one network and each machine has only one NIC, then that is not a problem. But in the case of multiple networks and/or NICs then it is necessary to pay attention to this detail.

3) Let's say "Node A" will be a transmitter -- then it is necessary to select a remote receiver node. Along the menu bar you'll see the label "Remote Receiver IP" and an IP address of 0.0.0.0. The star that moves from side-to-side is a visual cue that a remote receiver needs to be selected if this instance of NetStress is to act as a transmitter. Select the blinking '0.0.0.0' and the 'Select Remote Receiver dialog box will appear. Select the IP address of one of the remote receivers and press 'OK'.

4) Once a remote receiver has been selected, then the 'Settings' and 'Start' buttons along the top menu are activated.The 'Settings' dialog box allows you to control the following parameters:

•        TCP vs UDP
•        Number of data streams
•        TCP Packet Size
•        UDP Packet Size
•        Rate of Packet Transmission (Packets Per Second)
•        Data Flow
•        MTU
•        Display Units

5) While transmitting, certain buttons on the main menu change state -- e.g. the Stop button now becomes active. The Settings, Start, and 'Remote Receiver IP' buttons are temporarily deactivated.

6) To stop transmitting, press the Stop button.

7) Below the main menu there are two main windows -- 'Transmitter' and 'Receiver'. When NetStress is acting as a transmitter then the timecourse charts in the upper window will display the results. When NetStress is acting as a remote receiver then the lower window will display the results.

Copyright © 2010, Nuts About Nets, LLC