RT-SCALER: Adaptive Resource Allocation Framework for Real-Time Containers
Authors:
Vaclav Struhar, Silviu S. Craciunas, Mohammad Ashjaei, Moris Behnam and Alessandro Papadopoulos
Keywords:
container-based virtualization, real-time, resource adaptation
Abstract:
"Container-based virtualization has emerged as an advantageous deployment model in fog computing platforms since it enables the seamless co-location of applications in a heterogeneous environment with minimal overhead. For some application domains requiring a certain degree of predictability in the time domain (e.g., industrial automation), the adoption of container-based virtualization is not straightforward since the technology is not built to support real-time properties. In this paper, we propose RT-SCALER, which is a framework for adaptive resource allocation and dimensioning for real-time containers. RT-SCALER dynamically adapts the resource reservation of real-time enabled containers in order to improve the temporal predictability of the real-time applications running within the containers. We discuss the high-level orchestration approach, relating the different control levels, and give some practical insights into node-level container adaptation."