Nowadays, cloud-based services are becoming increasingly popular among service providers. As a result of virtualization these services run on non-dedicated hardware instead of just resources provided by multiple machines. In the cloud environment, these services run on virtual machines, which are managed by the open source OpenStack platform, where computing and storage resources are scalable. In real cloud environments these services usually run on distributed systems, thus controlling and managing network traffic is of utmost importance. The Java-based SDN (Software-Defined Networking) controller provided by the open source OpenDaylight platform is well-suited for such tasks. The controller provides the ability to dynamically control the incoming network traffic.
In the course of my thesis I study the OpenStack cloud platform, its components, and their functioning. I implement OpenStack, using the Fuel automation tool, in a virtualized environment where the OpenDaylight controller manages the OpenStack networks. The goal of my thesis is to construct a tunnel connection between two virtual machines running in different OpenStack environment, so that network funkctions can be moved between the OpenStack clouds on demand, increasing the flexibility and scalability of the service.
For cloud providers, it is important to measure and store network statistics, which can be used to determine network performance and utilization. In the course of my thesis, I using the OpenDaylight network controllers to monitor the traffic between virtual machines running in different clouds. I show the correct functioning of the two systems by running integration tests.