Container CPU Core Scheduling for Low-Interference Workloads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current container management systems, such as Kubernetes, fail to allocate CPU cores optimally to containers based on their specific needs, leading to uneven resource distribution, interference, and reduced performance, especially for latency-sensitive applications.
Innovation Solution
A method and device for scheduling resources that acquire parameters of containers, determine a designated core based on a scheduling policy, and allocate containers to the most suitable CPU cores, reducing interference and optimizing resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If containers are allocated CPU cores based on actual needs in current Kubernetes systems, then each container can obtain enough computing power, but resource distribution becomes uneven and interference occurs between containers
Solution Approach 1:
The patent segments CPU cores into different types (performance cores and efficiency cores) and assigns containers to specific core types based on their workload characteristics. This segmentation prevents interference between different container workloads by isolating them on appropriate core types, while still ensuring each container receives adequate computing power.
Solution Approach 2:
The patent applies local quality by matching specific container workloads to specific CPU core types. Latency-sensitive containers are allocated to performance cores with higher single-thread performance, while compute-intensive containers are allocated to efficiency cores. This localized optimization ensures each container runs on cores with qualities suited to its specific needs, reducing interference and improving overall system performance.
2Productivity
If containers are allocated CPU cores dynamically, then resource utilization improves, but response speed for latency-sensitive applications decreases due to suboptimal core selection
Solution Approach 1:
The patent performs preliminary action by pre-classifying CPU cores into different types (performance cores and efficiency cores) and pre-identifying the workload characteristics of containers. When allocating resources, the system quickly matches container types to appropriate core types based on pre-established criteria, avoiding the need for complex real-time decisions and ensuring latency-sensitive applications are rapidly assigned to performance cores for optimal response speed.
Data Source
AI summary
The subject technology relates to scheduling resources. For instance, an example method determines containers running on cores of a processor. The method further includes acquiring parameters of the containers. The method further includes determining a designated core from the cores that is adapted to a target container in the containers based on the parameters and a scheduling policy. The method further includes scheduling the target container in the containers to run on the designated core adapted to the container. In this way, containers can be scheduled to processor cores that are most suitable for their running, so that the response speed of services is increased and the mutual interference of resource contention between cores is reduced, thus improving the resource utilization and overall performance of the system.