VINES: Track 1: Enabling Low-Latency Applications Over a Best Effort Internet

The Internet often struggles to deliver data fast enough for real-time activities such as virtual reality, remote driving, and interactive medical operations. This problem occurs because the Internet typically sends data along a single path that can become congested, causing delays. This collaborative project brings together investigators from Columbia University and the University of Illinois at Urbana-Champaign to address these issues. By creating a system that sends information along multiple paths simultaneously, the project will ensure that even if one path is slow or unreliable, the data will arrive quickly, allowing high-performance applications to run smoothly for everyone. The intellectual merit lies in an end-to-end, cross-layer architecture that treats latency management as a core property of the end-to-end system, moving beyond traditional best-effort limitations. The research proposes to proactively send redundant information to mask Internet loss and delay, without relying on traditional reactive, slow recovery methods. Work is organized into specific thrusts, including developing a multipath-aware architecture to expose parallel wide-area routes, building cross-layer transport protocols to decide which paths and how much redundancy to use in real time, and integrating these solutions with emerging Internet standards for practical deployability. The cross-layer approach will allocate redundancy where it is most useful, such as for time-critical messages, to balance performance with overhead. By intelligently utilizing multiple diverse wide-area paths in parallel, the design reduces tail latency and jitter, providing a robust solution to the limitations of single-path best-effort delivery. The project has significant potential for impact by providing consistent low latency in a deployable method across the current Internet. This reliable performance can enable a wide set of new interactive applications and associated new markets, such as networked virtual and augmented reality, remote driving, and interactive medical applications, while also improving current applications like real-time videoconferencing. The project complements ongoing technological advancements in wireless connectivity to produce a complete end-to-end low-latency Internet service. In addition, the project will support the training of graduate students and will release code and non-proprietary data as open-source. 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.