Collaborative Research: ENTeR: Enabling NeTwork Research and the Evolution of a Next Generation Midscale Research Infrastructure

Over the past decade the Global Environment for Networking Innovations (GENI) [3, 2] successfully proved itself an effective platform for research and instruction across a range of CISE focus ar- eas such as: distributed computing and edge clouds, networking (software defined networking, hardware, software, wireless, mobile, local, regional, wide area), security, future clouds and smart cities. The successes resulted from several key accomplishments including: a federated approach for provisioning, accessing, and programming distributed cyberinfrastructure through standard abstractions, APIs, and security policies; a suite of shared services, like network stitching, ex- periment support and instrumentation; community-contributed resources - e.g., network con- nectivity, VLANs, bandwidth, compute racks; community outreach, education and instructional materials, e.g., labs, tutorials, videos, workshops. The successful GENI research story is captured in over 370 publications [4] by almost 800 authors, describing the science done in designing and implementing the testbed and in running many thousands of experiments on it that helped validate a broad range of scientific results. These experiments are continuing today, as GENI adds new users , running new experiments or using it as an important educational tool. The recent GENI experiments include work on evaluating elements of future exascale systems, securing scientific workflows, defining and testing novel SDX architectures, network security research, routing and transport protocol evaluations and many others . The strong community built around GENI is already discussing follow-on concepts of dis- tributed mid-scale experimental infrastructure that will help advance the state of the art in testbeds and advance the many aspects of computer science, while also contributing to research in other science domains [1]. For these reasons it is critical to continue running the GENI infrastructure and supporting the important research continuing on the testbed and allow the community to better understand the requirements for the next distributed experimental mid-scale infrastructure that will be needed to advance distributed systems, networking and domain research in the US. This proposal addresses an emerging research challenge - namely the need to develop a new, advanced, midscale platform for CISE research - by updating, adapting, and operating the exist- ing infrastructure for a two year time period that will both support ongoing novel at-scale network research experiments and, at the same time, enable the planning for a next generation midscale research infrastructure. Our proposed ENTeR project will facilitate the transition from GENI to a new platform capable of supporting a much larger portion of the CISE community. The EN- TeR team will consists of the University of Kentucky, RENCI/UNC Chapel Hill, University of Maryland, University of Utah and Internet2 - all intimately familiar with designing, developing, installing, and operating the various elements of the GENI infrastructure. The team's primary focus will be to ensure that the critical portions of the GENI infrastructure needed by the major- ity of the researchers remain available and relevant. Critical components include the InstaGENI and ExoGENI edge clouds, the stitching service components, which allow experimenters to build inter-site topologies, the clearinghouse, GENI portal, and GENI Desktop that provide easy ex- perimenter access to GENI tools, and the GENI monitoring service that allows users and opera- tors/providers to collect infrastructure measurement data. The team will also continue to perform various mission-critical management functions, like project and account management, support for tutorials and other outreach events, issue tracking, troubleshooting and escalation to appropriate responsible parties, including security incidents. The team will also participate in planning efforts for a follow-on midscale testbed infrastructure, adapting ENTeR project efforts in an attempt to begin moving toward any future infrastructure that may emerge [1]. The proposal describes the responsibilities of individual members of the team, the metrics by which the performance of the team will be evaluated and the interfaces which the team will maintain internally and with other communities in order to keep the infrastructure relevant for the proposed period of performance.