Dynamic Cloud Resource Scheduling Strategy Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud systems face challenges in efficiently scheduling computing, storage, and networking resources due to the use of a single static scheduling strategy, leading to sub-optimal or poor-quality resource allocation, inability to adapt to diverse workloads, and failure to leverage multiple scheduling strategies effectively.
Innovation Solution
A method and apparatus for dynamically selecting the most appropriate scheduling strategy from a set of candidate strategies based on available infrastructure resources and resource requests, using static characteristics analysis and speculative execution to optimize resource allocation, ensuring cost-effectiveness, speed, and scalability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single static scheduling strategy is used, then the system is simple to implement, but the resource allocation quality becomes sub-optimal or poor
Solution Approach 1:
The patent transforms the static scheduling strategy into a dynamic multi-strategy system that can adaptively select different scheduling algorithms based on workload characteristics. The system maintains multiple candidate scheduling strategies and dynamically chooses the most appropriate one for each scheduling scenario, thereby improving resource allocation quality without requiring complete system redesign.
Solution Approach 2:
The patent creates a universal scheduling framework that can handle multiple types of workloads and scheduling requirements through a single system architecture. This framework integrates multiple scheduling strategies (e.g., different algorithms for different workload types) and provides a unified interface for resource allocation, making the system both versatile and efficient.
2Productivity
If multiple scheduling strategies are implemented, then the resource allocation quality improves, but the system complexity increases
Solution Approach 1:
The patent introduces a scheduling strategy selection mechanism that acts as an intermediary between multiple scheduling algorithms and the resource allocation process. This mediator evaluates workload characteristics and selects the most appropriate scheduling strategy, thereby managing the complexity of multiple algorithms while maintaining high resource allocation quality.
Solution Approach 2:
The patent segments the scheduling process into distinct phases: workload characterization, strategy selection, and execution. By dividing the complex multi-strategy scheduling problem into manageable segments, the system can effectively handle multiple scheduling algorithms without being overwhelmed by complexity.
3Ease of operation
If a single scheduling strategy is used, then the system is easy to operate, but the adaptability to diverse workloads deteriorates
Solution Approach 1:
The patent implements a self-service scheduling system that automatically characterizes workloads and selects appropriate scheduling strategies without requiring manual intervention. The system monitors workload characteristics, evaluates them against predefined criteria, and autonomously chooses the best scheduling algorithm, thereby maintaining ease of operation while achieving high adaptability to diverse workloads.
Data Source
AI summary
Embodiments of the present disclosure provide a method and apparatus for a cloud system. The method includes: dynamically selecting, based on available infrastructure resources and at least one resource request, a scheduling strategy from a set of candidate scheduling strategies; and applying the selected scheduling strategy to schedule the infrastructure resources to serve the at least one resource request. Through embodiments of the present disclosure, when a single resource request or a batch of resource requests arrive, the most appropriate scheduling strategy is dynamically selected to generate an optimal allocation scheme for the request(s), thereby achieving cost-effective operations with the service level requirement of the resource request(s) satisfied.