Serverless edge computing has emerged as a transformative paradigm to support latency-sensitive applications, specifically smart applications such as Autonomous Vehicles (AVs). Although existing research works studied AV services scheduling, container retention and offloading techniques, critical challenges such as AV services’ images prefetching and AV mobility aware deployment over potential Roadside units (RSUs) needs to be addressed in detail. In this work, we propose Autonomous Vehicle Services Deployment and Scheduling Framework (AVSDSF [1]) to address AVs’ service requests handling. Proposed framework has been designed with the help of the following two algorithms. First algorithm, which runs at the Base Station (BS) to handle AV mobility and services aware images prefetching and deployment, is designed using Breadth First Search (BFS) and Depth First Search (DFS) approaches. It helps in minimizing cloud interaction latencies and AV request handling latencies. Second algorithm runs at the RSU and it handles the AV mobility and services aware scheduling to minimize AV request handling latency and optimally utilize various edge RSU resources. In extensive performance evaluations compared to existing approaches, our proposed framework is ensuring 47.47% higher AV services handling success rate, 23.99% higher success rate for high priority AV services, reduction of cloud communication latency by 35.79%. Moreover, compared to existing AV services distribution handling approaches, AVSDSF is ensuring 38.67% lesser AV service images distribution time and 59% fair utilization of edge RSU resources.
