Nowadays, where more than 85% of people have their own cell phones in the world, when an average person typically uses more than one smart IT tool daily, the demands for performance of telecommunication networks and the variety of new offered services are suddenly increased. In order to meet these expectations, service providers have to improve their networks and implement new services continuously which can be expensive and time-consuming processes. This is due to the fact that networks are currently static. It means that network elements - like firewall, cellular network controllers, etc.- are running on a hardware that is specially designed for this purpose. Using static networks results long product cycle, low resource utilization, slow development and expensive specialized hardware. To solve these problems, one possible dynamic solution is Network Function Virtualization (NFV) which aims to substitute current target devices and implement them in software. The advantage of NFV is that the software is applicable over general purpose hardware, furthermore the location of the software is changeable even while it is running. Networks and streams inside the hardware must be configurable at all time and have to be dynamically programmable, so the migration of the running software is possible from one hardware to another. These features are provided by Software Defined Networking. The application of SDN technologies offer many advantages, among other things it allows fast implementation of new services, and easy modification of existing ones.
The virtualized and software network functions are called Virtual Network Function. VNFs set specific compute requirements - such as CPU, RAM and storage - so they can be applied only on hardware that has enough free resources. While VNFs virtualize specific network elements, their sequential chains - called Service Chain (SC) - are applied for the end-to-end implementation of a service. These chains set operational criterias between the functions such as minimum bandwidth and maximal delay.
Network providers are aiming to serve as many users as possible, so they should be able to deploy a wide variety of services quickly and efficiently. There is a reference architecture for the management and orchestration of the NFV system (called Network Function Virtualization Management and Orchestration, NFV MANO), standardized by ETSI. The main purpose of NFV MANO is to provide rapid and flexible deployment and running of network functions over cloud systems and to avoid potential network congestion by monitoring network traffic.
Since the NFV systems are still in their infancy and the development is still ongoing process thus many questions may arise regarding their operation.
What kind of solutions exist today? How do they meet the ETSI reference architecture? How can they provide a big picture about networks and resources where the services are deployed? What determines where the VNFs are implemented? How can they optimally solve this challenge? During my diploma, I was looking for the answers to these questions above.