Task Scheduling Cycle Table Weighted Round Robin
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing task scheduling techniques, such as weighted round robin, often lead to unbalanced distribution of tasks among entities with varying capabilities, causing some entities to be overloaded while others remain underutilized, and do not effectively consider the individual abilities of entities in task assignment.
Innovation Solution
The implementation of a cycle table data structure or cycle array function that segments task allocations into cycles, ensuring only eligible entities with corresponding weight values receive tasks, thereby distributing tasks more evenly based on their relative capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If weighted round robin scheduling is used to distribute tasks to entities with different capabilities, then task distribution considers entity abilities, but some entities become overloaded while others remain underutilized
Solution Approach 1:
The patent segments the task distribution process into discrete cycles with multiple phases. Within each cycle, tasks are distributed in stages: first to entities with weight=1, then weight=2, and so on. This segmentation prevents any single entity from being overloaded while ensuring all entities are utilized appropriately according to their capabilities.
Solution Approach 2:
The patent implements periodic cycling through phases where the distribution weight increments and resets. Each cycle distributes tasks to entities with increasing weight values (1, 2, 3, ..., N), then resets to 1 for the next cycle. This periodic action ensures balanced utilization across all entities over time while adapting to their individual capabilities.
2Productivity
If tasks are distributed based on entity weights reflecting processing capacity, then higher capacity entities receive more tasks, but task allocation becomes unbalanced with bursty allocations
Solution Approach 1:
The patent segments task allocation into phased cycles where tasks are distributed to entities with weight=1, then weight=2, and so on through weight=N, before resetting to 1. This segmentation creates a systematic progression that smooths out bursty allocations while maintaining throughput efficiency by ensuring higher capacity entities receive appropriate task volumes.
Solution Approach 2:
The patent dynamically changes the distribution weight parameter across cycles, incrementing it from 1 to N and then resetting to 1. This parameter change strategy transforms the static weighted round robin approach into a dynamic system that alternates between favoring different entities, thereby smoothing allocations while preserving overall productivity.
3Device complexity
If all entities are considered to have equal capability as in round robin scheduling, then simple distribution is achieved, but entities with different processing capacities are not properly utilized
Solution Approach 1:
The patent introduces dynamics into the scheduling system by making the distribution weight a time-varying parameter that increments and resets across cycles. This dynamic approach allows the system to adapt to different entity capabilities without requiring complex real-time assessments, maintaining relative simplicity while improving capability utilization.
Solution Approach 2:
The patent changes the distribution weight parameter systematically across cycles, transitioning from weight=1 to weight=N and back to 1. This parameter change mechanism enables the system to account for varying entity capabilities without significantly increasing algorithmic complexity, as the weight progression follows a simple predictable pattern.
Data Source
AI summary
Mechanisms are provided for distributing work requests to worker devices. The mechanisms generate a cycle table data structure which segments a set of work request allocations into a plurality of cycles in which, for each cycle, a subset of worker devices, from a set of worker devices, are eligible to receive allocations of work requests. The mechanisms receive a work request from a computing device and select a worker device to receive the work request, from a first subset of worker devices that are eligible to receive allocations of work requests for a current cycle in the cycle table data structure, based on both of entries in the cycle table data structure corresponding to a current cycle, and execution of weighted round robin scheduling logic. The mechanisms distribute the work request to the selected worker device which performs an operation on the work request.