1. Create four compute nodes.
2. Create four VMs/compute node and an OVS instance that connects the VMs on a compute node.
3. Create an OVS instance to connect the four compute nodes.
4. Create traffic between different pairs of VM instances such that:
Measure the performance of OVS, in terms of the number of packets lost for UDP and throughput in the case of TCP connections. Observe the amount of traffic at which the OVS saturates if the VMs are on the same compute node and different compute nodes.
Tools such as iperf or netperf can be used to create traffic.
Finally, plot graphs of packet loss and throughput for different numbers of VMs communicating. Analyse the reasons for the performance and write a term paper describing the experimental setup, the results of the experimentation and the analysis of the results.