KuVS Summer School 2016
Software-Defined Networking and Network Function Virtualization
Hackathon cancelled!
Due to lack of interest, we had to cancel the ONOS hackathon.
That fixes the schedule of the summer school: We run from Monday morning to Thursday after lunch.
Welcome to the 2016 edition of the KuVS Summer School! June 20-23, 2016, Glücksburg |
Each year, the KuVS group organizes a Summer School on a hot topic of networking and distributed systems. The main target audience are first-year PhD students as well as people from industry.
KuVS - short for "Kommunikation und Verteilte Systeme", or Communication and Distributed Systems - is a technical committee of the German Society for Computer Science (Gesellschaft für Informatik) and the VDE's Information Society (Informationstechnischen Gesellschaft des VDE). It gathers interest in communication and distributed systems from both industry and academia and provides services to the community as well as represents these topics towards the general public.
Topic
The topic for the 2016 edition of the KuVS summer school is Software-Defined Networking and Network Function Virtualization.
Software-Defined Networking (SDN) has been a hot topic of networking research for the last several years. It centers on the idea of replacing the distributed control of typical networks by a centralized view, changing the way networks are controlled and operated. Typical application areas of SDN are, e.g., traffic engineering in both data centre networks and wide-area networks. SDN has created considerable interest from both academia and industry, resulting in a series of conferences on the topic as well as a range of products from practically every major networks provider.
Network Function Virtualization (NFV) started out from a very practical need: Network functions, e.g., firewalls, load balancers, or other kinds of middleboxes, where typically implemented as custom-tailored hardware, fulfilling one specific function. This entailed considerable disadvantages, for example long deployment times or insufficient flexibility for routing traffic. With the advent of high-performant virtualization techniques for network devices (and not only CPU or memory, as in conventional virtualization operating system techniques), it has become possible to supplant such custom-tailored hardware with general-purpose, commodity off-the-shelf hardware. This allows to run network functions at virtually any place in a network, turning it into a manageable and moveable commodity. For example, while so far traffic had to routed towards a place in the network where a particular network function was available, now it is possible to instantiate a network function where it is needed. Thinking this concept one step further gives the vision of "Distributed Cloud Computing" (or Edge Computing) where not only highly specialized network functions are considered, but arbitrary components (e.g., web servers) are commoditized in a similar fashion.
These two techniques - SDN and NFV - are tightly linked and complement each other. SDN is a useful technique to route traffic where it is conveniently used, NFV gives flexibility where and how traffic is to be processed. Both techniques will be necessary in future networks.
Program
To master either SDN or NFV requires a considerable skill set, from both theoretical and practical contexts. For example, much of the theoretical background rests in optimization techniques and graph-theoretical algorithms like facility location or multi-commodity flow problems. These concepts are necessary to exploit the full potential of SDN and NFV. From a practical side, experience with SDN controllers or NFV frameworks is to experiment with new algorithmic ideas in a concrete prototype or to push the boundary of these frameworks themselves.
It is the goal of this summer school to provide a comprehensive survey of the key skills necessary to successfully undertake independent research in SDN or NFV. We shall cover both theoretical backgrounds as well as practical aspects.
To this end, we create a program that rests on three main pillars:
- Presentations from leading experts with discussions and Q&A
- Hands-on work with actual frameworks
- Presentations by participants on their incipient or ongoing research work, along with feedbacks from other participants and experts.
The program will also encompass a one-day team building event to foster networking among the participants.
We are currently in the process of shaping the program details. See the program page for details, including a list of speakers who have tentatively agreed to participate.