Cloud Native Scheduler Plugins for Flexible Instance Preemption
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud native orchestration and scheduling systems lack flexibility and extensibility in their preemption mechanisms, as preset preemption rules cannot be easily added or deleted, leading to inflexible resource management.
Innovation Solution
Implement a plugin-based approach for judgment rules to determine instance preemption, allowing flexible loading and unloading of judgment rules, reducing coupling between the scheduler and judgment rules, and enabling easier rule configuration and management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If preset preemption rules are hard-coded in the scheduler, then the system structure is simple, but the flexibility and extensibility of preemption mechanisms deteriorates
Solution Approach 1:
The patent segments the preemption mechanism into independent judgment rule plugins that can be separately developed, configured, and loaded. Each judgment rule is encapsulated as a discrete plugin module that evaluates specific preemption conditions (e.g., instance priority, resource usage, QoS requirements), allowing the system to maintain a simple core scheduler while achieving high flexibility through modular rule components.
Solution Approach 2:
The patent implements dynamic rule loading and configuration capabilities, allowing judgment rules to be added, removed, or modified at runtime without restarting the scheduler. The system dynamically selects and applies appropriate judgment rules based on current scheduling scenarios, enabling the preemption mechanism to adapt to changing requirements while maintaining system simplicity.
2Adaptability or versatility
If multiple judgment rules are integrated into the scheduler, then the preemption mechanism becomes more flexible, but the device complexity increases
Solution Approach 1:
Each judgment rule is encapsulated as an independent plugin with a standardized interface, separating the complexity of multiple rules from the core scheduler. The scheduler maintains a simple rule registry and selection mechanism, while the actual rule logic resides in separate, manageable plugin modules that can be independently developed and maintained.
Solution Approach 2:
The patent designs a universal plugin interface that all judgment rules implement, allowing diverse preemption rules to be managed through a common framework. This universal interface handles rule registration, configuration, and execution, enabling the scheduler to support multiple judgment rules without increasing its own structural complexity.
3Ease of operation
If preemption rules are easily configurable, then the ease of operation improves, but the system requires more complex plugin management infrastructure
Solution Approach 1:
The patent implements self-service mechanisms where the scheduler automatically discovers, loads, and configures judgment rule plugins based on predefined criteria. The system automatically manages plugin lifecycles including activation, deactivation, and version management, reducing the burden on operators while providing easy configuration capabilities through standardized interfaces.
Data Source
AI summary
Embodiments of the present disclosure provide a scheduling method and device for a cloud native orchestration and scheduling system and a storage medium. A target judgment rule plugin is determined from multiple preset judgment rule plugins included in a preset judgment rule plugin set and loaded into a scheduler of the cloud native orchestration and scheduling system; whether any instance on any node in the cloud native orchestration and scheduling system is allowed to be preempted by a to-be-deployed instance is sequentially judged by using the target judgment rule plugin for the to-be-deployed instance; a candidate to-be-preempted instance on a candidate to-be-preempted node is determined; candidate to-be-preempted nodes are sorted by using a preset comparison plugin in the scheduler to determine a target to-be-preempted node, and the to-be-deployed instance is controlled to preempt the candidate to-be-preempted instance on the target to-be-preempted node.