Performance Optimization of Control Applications on Fog Computing Platforms Using Scheduling and Isolation

Abstract

The research leading to these results has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 764785, FORA-Fog Computing for Robotics and Industrial Automation.

Communication Scheduling for Control Performance in TSN-Based Fog Computing Platforms

“…An FCP hosts multiple applications of mixed-criticalities, e.g., critical control applications, real-time applications, and best effort applications. Applications are typically modeled as interacting periodic real-time tasks that exchange messages, see [17] for how application tasks can be modeled. In this paper we address the configuration of the TSN communication infrastructure, hence we focus on messages.…”

“…Researchers have proposed several ways of putting together the schedules for tasks and messages in a global system configuration, e.g., by combining the formulation of their scheduling problems [24] or by iteratively integrating the task and message scheduling. The solution presented in this paper for flows can be combined with the formulation for tasks from [17]. In addition, to support the integration of the GCLs that we determine with tasks schedules derived separately, we maximize the time duration where tasks have to execute, denoted with E in Fig.…”

“…We have proposed a Constraint Programming (CP)based solution for deriving the ST communication schedules. We have addressed the problem of scheduling of tasks on an FCP for QoC [17], which is orthogonal to the message scheduling problem. However, to facilitate the integration of tasks and message schedules, our CP implementation also aims at supporting the integration of tasks and messages by creating space in the communication schedule timelines, where tasks need to execute.…”

Mapping and Scheduling Automotive Applications on ADAS Platforms using Metaheuristics

“…We describe a Simulated Annealing (SA)-based metaheuristic approach, first introduced in [6], [7] and extended in [28], [29], which uses an EDF-based heuristic to solve the task scheduling problem. The scheduling heuristic allows task preemption by simulating an Earliest Deadline First (EDF) scheduling policy parameterized by task offsets and local deadlines decided by SA, see Section IV-C. For a mapping of task to cores we assign, according to our SA strategy, an offset and a deadline to each task and simulate EDF to obtain the static schedule, if one can be found.…”

Fog Computing Based on Machine Learning: A Review

“…However, at a higher level, fog nodes can handle frameworks like Weka [31] and Scikit-learn to implement many AI applications. ML is used to execute, optimize, assign, or monitor functional tasks such as clustering, routing, duty-cycle management, data aggregation, and medium access control [32]. It is not easy to manage the relevant processes in the fog nodes because they are dynamic, complex, and heterogeneous.…”