Event-Based Task Scheduling for Modular Frameworks Without Polling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional task scheduling in modular extensible frameworks relies on inefficient greedy and thread-poll multi-thread schedulers that consume significant computational resources by continually checking node status, leading to suboptimal performance.
Innovation Solution
Implementing an event-based task scheduling system that utilizes a dispatcher to monitor node events and conditions, such as task completion and availability, rather than continuous status checks, to improve scheduling efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If continuous status checks are used to monitor node availability, then task scheduling reliability is improved, but CPU usage increases significantly
Solution Approach 1:
The patent replaces continuous polling with periodic event-driven notifications. Nodes notify the scheduler of status changes (task completion, availability) through events rather than continuous checks. This periodic action occurs only when state changes, significantly reducing CPU usage while maintaining scheduling reliability.
Solution Approach 2:
Nodes autonomously generate and send event notifications when their status changes. Instead of the scheduler actively querying each node continuously, nodes self-serve by proactively informing the scheduler of their state, reducing the computational burden on the scheduling system while ensuring timely status updates.
2Measurement precision
If continuous status checks are performed to ensure accurate node availability, then scheduling precision is improved, but computational resource consumption increases
Solution Approach 1:
Status monitoring transitions from continuous periodic checks to event-triggered periodic updates. The system maintains precise node status detection by receiving notifications only when actual state changes occur, eliminating redundant checks and reducing computational resource consumption while preserving measurement precision.
Solution Approach 2:
The system implements feedback through event notifications where nodes provide status information to the scheduler only when changes occur. This feedback mechanism ensures the scheduler has accurate, up-to-date node availability information without the energy cost of continuous polling, as updates are triggered by actual state changes rather than fixed intervals.
3Speed
If event-based notifications are implemented for task completion, then task advancement speed is improved, but system complexity increases
Solution Approach 1:
The patent introduces an event notification mechanism as an intermediary between node task completion and scheduler awareness. Events act as mediators that automatically convey status changes, enabling faster task advancement without requiring complex continuous monitoring logic. The intermediary simplifies the interaction model while improving response speed.
Solution Approach 2:
The system replaces the mechanical polling approach with an event-driven notification system. Instead of the scheduler mechanically checking each node's status continuously, events are generated and propagated automatically when state changes occur. This substitution reduces system complexity by eliminating the need for complex polling coordination while improving task advancement speed through immediate notification.
Data Source
AI summary
Disclosed are apparatuses, systems, and methods for software-agnostic retrievals of event-based task scheduling. The systems and methods utilize a dispatcher to monitor events within the system to provide the scheduler with nodes ready for execution. Using an indicator, the system can determine, that the first node assigned to a first phase of a task is in a first state of a plurality of states. Based on the first state, the identifier can be added to a queue. A second phase of the task can be determined to be in a second state, the first identifier may be provided to a scheduler to advance the task to the second phase.