This research will generate new knowledge about the design and deployment of resilient networks using GENI. It will evaluate alternatives for future Internet design using the GpENI infrastructure, under the GENI PlanetLab control framework, and -- stitched together with several ProtoGENI facilities -- perform resilience and survivability experiments at scale, investigating, in particular, the geographic scope needed to emulate large-scale disasters. GpENI provides the ability to construct a large set of arbitrary backbone topologies, while the dense ProtoGENI clusters will emulate both wired and wireless access networks. These experiments will cross-verify analytical and simulation-based resilience research currently underway, leveraging topology and challenge generation tools developed for this purpose, and emphasizing multi-path multi-realm diverse transport protocols. The broader impact of the project includes fulfilling the vision of GENI supporting future internet design, and contributing to the education of graduate students at Kansas Univeristy.
This goal of this project was to use a large-scale network testbed to perform experiments not possible on small in-laboratory testbeds. GpENI (Great Plains Environment for Network Innovation) is a project funded by the NSF GENI (Global Environments for Network Innovation). GENI funds programmable network testbeds at a large scale so that network researchers can try and validate new ideas without affecting the commercial networks that make up the Internet. Unfortunately, there are many testbeds that are underused, so the NSF has funded projects such as this one to show the utility of the GENI testbed for running experiments. GpENI is a global programmable Future Internet testbed centered at The University of Kansas that provides deep network programmablity, including the ability to change topologies and emulate nework failures from disasters or attacks against the network infrastructure. Therefore, in this project we use these capabilities to experimentally analyze the resilience of networks to node failure and how well diversity mechanisms perform that we designed in our NSF PostModern Internet project. Ideally, new architectures and algorithms should be cross-verified among several different methodologies to give higher confidence in their analysis. For this project, we show that is it possible to use a large-scale experimental testbed to validate diversity mechanisms that have been verified by analytical and simulation methods. We experimentally validate the robustness of alternative topologies to node failures in the network. In addition to the publications from this project, we publicly demonstrated the experiments at the 17th GENI Engineering Conference (GEC) in Madison WI in July 2013. In addition to contributing to the education of several graduate students, the broader impacts of this project include contributing toward a more resilient Future Internet. The resilience of the Future Internet against attacks and large-scale disasters is of major importance to society, which relies on this critical infrastructure.
