Cluster Scheduling Snapshot Updates Using Incremental Node Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cluster schedulers in large-scale clusters face significant calculation losses during snapshot updates and preemption processes, leading to crashes and impaired scheduling efficiency.
Innovation Solution
Implement a cluster-based scheduling method that utilizes a historical scheduling snapshot and a scheduling cache to selectively update node information based on chronological timing values, employing incremental synchronization and balanced binary search trees to optimize snapshot updates and preemption processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all node information is synchronized to scheduling snapshot during each scheduling, then scheduling decision accuracy is improved, but calculation loss increases and scheduler stability deteriorates
Solution Approach 1:
The patent extracts only the changed node information from the scheduling cache and applies incremental updates to the scheduling snapshot, rather than synchronizing all node information. This is achieved by comparing timing values to identify changed nodes and updating only those specific nodes in the snapshot, thereby reducing calculation loss while maintaining scheduling accuracy.
Solution Approach 2:
The patent performs partial synchronization by updating only the necessary portion of the scheduling snapshot (changed nodes) rather than the entire snapshot. The incremental update mechanism applies partial actions selectively based on timing value comparisons, reducing overall calculation overhead while preserving essential scheduling information.
2Measurement precision
If all node information is synchronized to scheduling snapshot during each scheduling, then scheduling decision accuracy is improved, but scheduler stability deteriorates due to frequent crashes
Solution Approach 1:
The patent extracts only the changed node information from the scheduling cache and applies incremental updates to the scheduling snapshot, rather than synchronizing all node information. This is achieved by comparing timing values to identify changed nodes and updating only those specific nodes in the snapshot, thereby reducing calculation loss while maintaining scheduling accuracy.
Solution Approach 2:
The patent prepares the scheduling snapshot in advance with incremental update capabilities and timing value metadata. By pre-structuring the snapshot to support selective updates and having the timing value comparison mechanism ready, the system cushions against the computational burden that would otherwise cause scheduler crashes, improving stability.
3Productivity
If incremental update based on timing values is used, then calculation loss is reduced and scheduling efficiency is improved, but device complexity increases
Solution Approach 1:
The patent performs preliminary actions by maintaining timing values for each node in the scheduling cache and recording the latest change timing in the scheduling snapshot before the actual update occurs. This pre-preparation of timing metadata enables efficient incremental updates without requiring complex real-time analysis during scheduling decisions.
Solution Approach 2:
The patent introduces timing values as an intermediary mechanism between the scheduling cache and the scheduling snapshot. These timing values act as a mediator that enables the system to identify changed nodes and perform incremental updates, simplifying the update logic while improving scheduling efficiency.
Data Source
AI summary
The present disclosure provides a cluster-based scheduling method, a medium and an electronic device, the method includes: in response to a current scheduling, obtaining a historical scheduling snapshot generated in a previous scheduling and reading a global timing value corresponding to a cluster during the previous scheduling from the historical scheduling snapshot, where the global timing value is a latest change timing of a node in the cluster during the previous scheduling recorded in chronological order; reading a respective timing value corresponding to each node in the cluster during the current scheduling from a scheduling cache and selecting, using the global timing value as a reference, a target node whose timing value is later than the global timing value in chronological order; and updating the historical scheduling snapshot based on node information of the target node to obtain a scheduling snapshot of the current scheduling.