FAI: Towards a Computational Foundation for Fair Network Learning

Both today’s cloud services and emerging IoT applications alter the point-to-point communication paradigm of the existing IP Internet architecture, and require increasingly demanding quality-of-service (QoS) requirements. These requirements call for scalable and intelligent network algorithms for controlling and coordinating various network components and managing and optimizing resource allocations, with capabilities 1) to meet ever stringent availability, reliability and QoS requirements demanded by emerging services; 2) to cope with the enormous complexity of networked systems; 3) to effectively exploit the rich diversity and redundancy inherent in such complex systems as well as the new capabilities offered by new networking architectures and technologies.This project puts forth a three-plane view of networking as a conceptual framework to structure network functions and guide us in the network algorithmic designs for timely, resilient and resource-efficient information delivery: 1) an information plane capturing application semantics and requirements; 2) a (logically) centralized control plane; and 3) a distributed (programmable) communication (data) plane. This project postulates two design principles and challenges in network algorithms: a) the need for co-design of centralized and distributed network algorithms that can take advantage of a centralized control plane with a global view of the network state, while also enabling the distributed (programmable) network elements to make fast and intelligent decisions to adapt to the changes in the network conditions (e.g., failures); and b) the need for just-in-time (near) optimality as a key metric to gauge and guide the design of network algorithms for timely, resilient and resource-efficient information delivery.