Software Defined Networks (SDN) is seen as the most promising solution to resolve the challenges in realizing sophisticated network control. SDN builds its promise on the separation of the network control functions from the network switching elements. By moving the control plane out from the network elements into standalone servers, the switching elements can remain simple, general-purpose, and cost-effective and at the same time the control plane can rely on design principles of distributed systems in its implementation instead of being confined to distributed routing protocols.

The purpose of the seminar was to look at the current developments in this quickly evolving problem domain and identify future research challenges. The seminar brought together researchers with different domains and backgrounds. Given the high level of interest in SDN from industry, the organizers also invited many participants from companies working with SDN related networking products and services. This mix of people resulted in fruitful discussions and interesting information exchange. The structure of the seminar took advantage of these different backgrounds by focusing on themed talks and group discussions.

Organization of the Seminar

Software-Defined Networking (SDN) continues to remain relevant both for the industry and academia and indeed this was very much reflected in the backgrounds of the seminar participants; the seminar had a balanced mix of representatives both from industry and academia.

These two very active communities, industry and academia, are pursuing SDN with different mindsets, different solutions and differection implications in mind, however. The organizers felt that the interactions had been clearly insufficient in the past: practical challenges in SDN continue to remain little known in the academia whereas the industry often remains unaware of the recent useful developments in research. To this end, the two and half day seminar was explicitly structured around this observation; the goal was to allow for fruitful interactions between the industry and academia to maximize the exchange of ideas, challenges and lessons learnt between these two communities.

The seminar discussions and talks were structured around three themes:

1. Status updates. From the very definition to the ongoing standardization work, SDN is still evolving. In these talks and discussions, we dived into the ongoing work at ONF as well as the perceived hard problems to be solved.
2. Industry use cases. In this theme the focus was on exposing the academia to the practical use cases on which industry is working.
3. Implementation. The third theme dived into the details and exposed the seminar participants to both the practical implementation issues faced as well as more theoretical observations about the system design.

The seminar was well received by the participants. Among the participants there were also organizers of future SDN workshops (IRTF SDN and DIMACS SDN) who signaled the intent of building their workshops around the similar discussion-oriented structure preferred at Dagstuhl.