Managing Bandwidth: The Key to Cloud-Assisted Autonomous Driving

TL;DR

This paper explores how cloud computing can be effectively used for real-time autonomous driving tasks by managing bandwidth to meet strict latency requirements, challenging traditional views.

Contribution

It proposes a framework for allocating bandwidth in cloud-assisted autonomous driving to ensure latency SLOs are met while maximizing system benefits.

Findings

Bandwidth management can enable reliable cloud-assisted autonomous driving.

Proper resource allocation balances latency and computational benefits.

Cloud offloading is feasible for time-critical autonomous driving tasks.

Abstract

Prevailing wisdom asserts that one cannot rely on the cloud for critical real-time control systems like self-driving cars. We argue that we can, and must. Following the trends of increasing model sizes, improvements in hardware, and evolving mobile networks, we identify an opportunity to offload parts of time-sensitive and latency-critical compute to the cloud. Doing so requires carefully allocating bandwidth to meet strict latency SLOs, while maximizing benefit to the car.