Towards Network Programmability: Network Function Virtualization and Software Defined Networking
|Instructor||Paganelli, Ronga (CNIT)|
|20/02/2017||09:00–13:00||108 Santa Marta (Firenze)|
This course aims at introducing two emerging paradigms, namely Software-Defined Networks (SDN) and Network Function Virtualization (NFV), which are expected to revolutionise the design, building, and management of networks. The main idea of Network Function Virtualization is implementing network functions as software components executed in virtual machines or containers provisioned in general-purpose hardware systems.
Thanks to a more agile lifecycle management of virtual machines, VNFs can be instantiated, updated, deleted when and where needed, as well as dynamically combined to provide more complex capabilities on demand. The main concept of SDN consists in decoupling the control plane (e.g., routing decision functions) from the hardware-based data plane (i.e., packet forwarding devices). Thanks to the clear separation of control and data forwarding functions, the control logic is placed in a logically centralized software (SDN Controller) which collects information from the forwarding devices to maintain an up-to-date topological representation of the network and provides an abstract view of the network to applications (traffic engineering, information-centric networking, etc.).
Enabled by a centralized control plane, innovative concepts can be implemented such as Cognitive Networking, where the OODA (Observe Orient Decide Act and learn) cognitive loop can be applied to networks in order to learn and react to unplanned events.
The course will introduce the main concepts of NFV and SDN, describing main reference architectural models (standard specifications) and implementations (open source software). An introduction of research challenges in the NFV/SDN will then close the course, discussing also possible interdisciplinary research topics (e.g. optimization algorithms, formal methods, cognitive networking, etc.).