This project funds an October 8, 2013 joint workshop between the Open Networking Foundation (ONF), GENI, and US Ignite communities. The goal of the workshop is to jointly explore the interaction of emerging uses of Software Defined Networking (SDN) in the US Ignite and GENI communities and SDN standardization roadmaps and feature sets. As a result, SDN is likely to become a more powerful tool for GENI network research experimentation, US Ignite pre-commercial applications, and commercial uses.
Communication between the communities will facilitate development of SDN standards and applications. Additionally, this project supports the involvement of 20 students in the workshop. US Ignite will seek two types of student participants: 1) students who are familiar with OpenFlow and working in Computer Science systems and networking and GENI or US Ignite projects; and 2) students who have been participating in US Ignite, Mozilla-Ignite, and GENI projects and can share information on application patterns of SDN use. Output of the workshop will be shared with the larger community via a workshop report posted to the US Ignite web site as well as other web pages, wikis and communication mechanisms.
US Ignite is engaged in the development and promotion of new gigabit applications for compelling new public benefit applications. These gigabit applications leverage three advanced technologies – gigabit transmission to end users, local clouds and Software Defined Networking (SDN). SDN, and in particular OpenFlow is a programmable network protocol that manages and directs traffic among routers and switches from various vendors, and separates the programming of routers and switches from the underlying hardware. While OpenFlow use in the research datacenter is well established, the US Ignite and GENI communities are exploring OpenFlow to federate distributed local clouds for its gigabit applications. The Open Networking Foundation (ONF) manages the standardization of OpenFlow protocols. ONF had expressed interest in the application usage patterns for OpenFlow, and wanted to gather information about emerging applications and their requirements for SDN capabilities. Likewise, the US Ignite and GENI communities have an interest in shaping future versions of OpenFlow to fit and suit their applications and experiments. US Ignite wanted to gather the latest information on the evolution of the OpenFlow protocol and SDN architectures. On October 8, 2013, US Ignite and ONF hosted a joint workshop in Sunnyvale, California to explore opportunities for collaboration in the development, deployment, and promotion of applications enabled by SDN, gigabit transmission, and local cloud technologies. Seventy-one students, professors, researchers and technologists from academia, research labs, and industry were in attendance. The Workshop was a unique opportunity for ONF to meet with gigabit application developers working on end-user applications and with researchers using SDN and Open Flow technology. (All previous guidance to the ONF had come from its carrier members.) ONF explained its mission to drive open SDN protocols. ONF emphasized the importance of its democratic governance over the SDN standards process. ONF also explained its progress on OpenFlow version 1.5, and other key initiatives such as building the OpenFlow substrate, fostering implementation and deployment, and encouraging SDN innovation. They also announced for the first time that they would attempt to standardize "usage profiles" called "table type patterns" or TTPs for the use of OpenFlow so that industry players could optimize their OpenFlow implementations around these usage profiles. GENI representatives explained how the infrastructure (i.e., networks and racks) are used for experimentation and how multiple experiments can be run concurrently using isolated infrastructure "slices" throughout the U.S. The US Ignite and GENI communities emphasized the importance of dynamic slicing and virtual networks. The US Ignite and GENI community is looking for slice support, native isolation between "slices", better performance, faster adoption of new standards (such as IPv6), ability to run different OpenFlow versions in the same hardware switch, and natively share the same physical OpenFlow switch. They also pointed out that although ONF was working on later OpenFlow protocol versions, many researchers are only able to run OpenFlow 1.0 or 1.1 on the GENI network. The participants discussed the limitations of open dialogue that ONF faces while managing the intellectual property of the protocols. Participants agreed that everyone is in a learning phase, that it is necessary to see what is actually used, and that agreements may be necessary on a case-by-case basis depending on the intellectual property involved. The participants also discussed that if US Ignite and GENI researchers can nominate a set of commonly used applications in the end-user SDN community, then ONF could make a matched effort on testing and interoperability against the set of applications and establish appropriate TTPs. It would help ONF arrive at requirements if use cases and applications (or at least their requirements) could be provided to ONF so they could take them as input and trials. As a result, the OpenFlow standard may become a more powerful tool for network research experimentation by GENI, pre-commercial applications by US Ignite, and commercial uses by developers following industry standards. The workshop was organized and moderated by Glenn Ricart of US Ignite and Dan Pitt of the ONF. At the conclusion of the workshop, Dan agreed on behalf of ONF to accept a member of the US Ignite and GENI community as an at-large member of the ONF. Further, the TTP committee of the ONF agreed to examine the table type patterns involved in a few of the US Ignite and GENI applications presented for possible standardization. Prior to this workshop, leadership in SDN and OpenFlow was divided between the research world and industry. But the information exchanged in this workshop created new relationships to help each sector understand and to begin building their mutual interests into their respective planning efforts. Everyone recognized the need for additional joint educational efforts regarding SDN – acknowledging the important broader impact on mutual collaboration to advance SDN technologies and protocols for the public benefit.
