CSR: SMALL: Foundations of Software-Defined Real-Time Networking

Washington State University, Pullman WA

Investigators

Abstract

Communication flows in cyber-physical systems (e.g., aircraft, automobiles, industrial systems) have stringent timing and performance constraints. Late delivery of important messages can result in safety violations or system failure. For instance, critical components of a car, such as the anti-lock braking module, may fail to function properly if they do not receive timely signals from the brake pedal through the communication bus—this puts the passengers’ safety at risk. These networks are carefully engineered to ensure resource requirements are met at runtime. This is a highly complex process since it is up to the designers of such systems to ensure correct and reliable network-wide (global) behavior, even in the face of failures. This project aims to ease this process and devise techniques for more efficient utilization of network resources, providing better end-to-end quality-of-service and resiliency guarantees. The results from this project will make many critical cyber-physical systems of modern society more resilient and, hence, much safer. Modern complex safety-critical networks, such as those used in avionics, automobiles, and power grids, need more sophisticated solutions to ensure that their end-to-end service guarantees (i.e., timing and data rate requirements) are met. The techniques developed for more general-purpose and data-center networks are not readily adaptable to safety-critical cyber-physical networks with stringent “real-time” requirements. The key innovations of this proposal involve creating holistic models, algorithms, and software designs to achieve the scalability, adaptability, and resiliency required for real-time applications. The proposed work will explore design methodologies for critical networks by combining real-time scheduling theory with new and emerging networking paradigms (viz., software-defined networks and programmable switches). Specifically, this foundational research devises novel routing, flow scheduling, and fault tolerance techniques for real-time networks, resulting in better admissibility and resilience, thus enhancing overall resource utilization. The project has two related research thrusts. The first thrust ensures end-to-end quality-of-service guarantees. Due to their safety-critical nature, real-time networks must be resilient to failures by design. Communication flows must reach their destinations even when there is a network failure. The second thrust addresses this aspect and devises resilient routing schemes. The mathematical frameworks constructed in this research for analyzing such systems will provide strong guarantees, potentially facilitating faster testing, integration, and verification efforts. The proposed solutions will be evaluated and tested on various platforms, from simulation engines to real hardware testbeds. All data, insights, results, and implemented frameworks will be publicly available online. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.