Uniform Execution Schedule Algorithm for Dynamic Workloads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scheduling algorithms face inefficiencies and potential irregularities when dynamically adjusting the execution schedule of periodic workload instances, especially when the number of instances changes at runtime, leading to suboptimal resource usage and potential delays in workload execution.
Innovation Solution
A globally uniform-in-time execution schedule algorithm that gradually adjusts execution start times to maintain uniformity, allowing for quick convergence to a globally uniform state upon additions or removals of workload instances, without requiring a complete schedule recalculation, thereby minimizing resource usage and ensuring timely execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the schedule is recalculated completely whenever workload instances are added or removed, then global uniformity is maintained, but computational complexity and time consumption increase significantly
Solution Approach 1:
The patent segments the schedule adjustment process into two distinct phases: (1) a complete recalculation phase that occurs only when the number of workload instances changes, and (2) a gradual convergence phase that adjusts individual execution start times incrementally toward their optimal uniformly-spaced positions. This segmentation avoids the need for complete recalculation after every minor change while still achieving global uniformity.
Solution Approach 2:
The patent introduces dynamic adjustment of execution start times where the schedule evolves gradually from its current state toward the optimal uniformly-spaced state. Instead of static complete recalculation, the system dynamically adjusts start times incrementally, allowing the schedule to adapt flexibly to changes in workload instance count while minimizing disruption to ongoing executions.
2Manufacturing precision
If the execution start times are adjusted frequently to maintain global uniformity, then schedule accuracy is improved, but resource usage and system overhead increase
Solution Approach 1:
The patent applies partial action by adjusting only the necessary execution start times that need correction, rather than recalculating and adjusting the entire schedule. The system identifies specific start times that deviate from uniform spacing and adjusts only those, consuming minimal computational resources while maintaining schedule accuracy.
Solution Approach 2:
The schedule adjustment mechanism serves itself by automatically detecting when global uniformity has been compromised and initiating minimal necessary adjustments. The system monitors its own state and performs self-correction without external intervention, reducing overhead by acting only when and where needed.
3Stability of the object's composition
If the schedule adjusts gradually to maintain stability, then system stability is improved, but convergence to uniform state takes longer
Solution Approach 1:
The patent performs preliminary adjustment by calculating the optimal uniformly-spaced execution start times in advance when the number of workload instances changes. This preliminary calculation establishes the target uniform state before gradual convergence begins, providing a clear destination for the incremental adjustments and enabling faster convergence compared to blind gradual adjustment.
Data Source
AI summary
An algorithm for efficiently maintaining a globally uniform-in-time execution schedule for a dynamically changing set of periodic workload instances is provided. At a high level, the algorithm operates by gradually adjusting execution start times in the schedule until they converge to a globally uniform state. In certain embodiments, the algorithm exhibits the property of “quick convergence,” which means that regardless of the number of periodic workload instances added or removed, the execution start times for all workload instances in the schedule will typically converge to a globally uniform state within a single cycle length from the time of the addition/removal event(s) (subject to a tunable “aggressiveness” parameter).