Towards Low-Latency Function-as-a-Service: Ensuring Real-Time Performance in Open Source FaaS

Function-as-a-Service (FaaS) is a highly efficient paradigm in serverless cloud computing, offering automated resource provisioning, scalability, simplified software and maintenance. Emerging applications in 5G and 6G networks, the Internet of Things, and Industry 4.0/5.0 increasingly necessitate the offloading of time-critical functions to cloud and edge computing environments. However, current FaaS platforms struggle to meet strict latency requirements for real-time applications. This study proposes an open-source prototype for a real-time FaaS system that supports both real-time and non-real-time functions, prioritizing time-critical workloads. Using simulated balancing robots with PID controllers, we evaluate performance under various scheduling strategies, including Completely Fair Scheduling and Earliest Deadline First. Results show that the proposed framework effectively supports real-time execution with improved resource utilization and reduced deadline misses.