Disruptions occur often even in well-managed wired networks due to various failures and more so in emerging wireless networks due to additional causes such as external interference. With the increased use of the Internet for many real-time business applications, there is a growing demand for high service availability despite such disruptions. The goal of this project is to develop resilient routing schemes that mitigate the impact of disruptions on mission-critical applications and services. Specifically, this project aims to: (i) devise a failure inferencing based fast rerouting scheme to provide failure protection similar to MPLS fast-reroute without altering the destination-based forwarding paradigm prevalent in IP networks; (ii) design scalable localized on-demand link state routing schemes that ensure reliable delivery in spite of disruptions without requiring accurate state at each node; (iii) explore ways to utilize both topology and position information for handling disruptions; (iv) extend and apply these methods for load balancing and resilient multicasting. This project will boost the robustness of the Internet and enable services that improve the lives of everyone. It also has significant educational value in addition to facilitating knowledge sharing and distance education as a result of an always-available Internet. The results from this project will be assimilated into advanced graduate level and introductory undergraduate level theory courses as well as hands-on laboratory courses. Furthermore, the research advances will also be disseminated to the industry and the broader community through several publications and prototypes.
The goal of this project is to make wired and wireless networks more resilient to failures and disruptions. Towards that goal, this project has designed and developed various routing schemes for failure resilience. In particular, a failure inferencing based fast rerouting approach is developed that provides high availability without jeopardizing network stability. It offers near-continuous forwarding of packets despite failures by initiating local rerouting as soon as a failure is detected. This approach ensures loop-free forwarding to a destination if there exists a path to it and thus protects an IP backbone network against any single link or router failure with no changes to the forwarding plane and minimal changes to the control plane of the Internet. A scalable localized on-demand link state routing approach is also designed for wireless mesh networks where router adjacencies are relatively static whereas link qualities are quite dynamic. This approach exploits the characteristics of mesh networks by propagating the state of a link only when needed and as far as necessary. It guarantees packet delivery to reachable destinations regardless of the extent of disruptions in the network without requiring accurate state at each router. Together, these schemes mitigate the impact of failures and disruptions on end users and enhance the network availability to the level suitable for supporting mission-critical applications and services. In terms of human resource development, this project supported the dissertation work of five PhD students and thesis work of three MS students. Four undergraduate students were also involved in this project and benefitted from the research experience.
