Distributed Resource Scheduling via Coordinator Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing distributed resource systems face challenges in efficiently scheduling user requests across multiple resource-providing entities, leading to performance issues due to resource competition and the inability to implement global scheduling without relying on user agents.
Innovation Solution
A method and apparatus for scheduling user requests in a distributed resource system, which includes schedulers and a coordinator, where the scheduler acquires resource consumption data from the coordinator and uses a scheduling algorithm to compute virtual start and finish times based on resource weights and consumption quantities, allowing for global scheduling and performance optimization without relying on user agents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If multiple schedulers independently allocate resources to user requests, then each scheduler can operate autonomously, but resource competition occurs and global scheduling optimization cannot be achieved
Solution Approach 1:
The patent introduces a coordinator as an intermediary component that collects resource consumption statistics from multiple schedulers and distributes them to relevant schedulers. This mediator enables global scheduling optimization without compromising the autonomy of individual schedulers, as the coordinator handles the complex inter-scheduler coordination while schedulers continue to operate independently using the provided statistics.
Solution Approach 2:
The patent implements a feedback mechanism where schedulers report resource consumption statistics to the coordinator, which then distributes this information back to schedulers. This feedback loop enables schedulers to make informed scheduling decisions based on global resource utilization patterns, improving overall resource efficiency while maintaining operational autonomy.
2Ease of operation
If user agents are used to collect request statistics and determine schedulers, then scheduling decisions can be made, but the system cannot implement global scheduling across multiple resource-providing entities
Solution Approach 1:
The coordinator acts as an intermediary that aggregates resource consumption statistics from multiple resource-providing entities and makes global scheduling decisions. This allows the system to achieve global scheduling capability without requiring user agents to have direct access to or knowledge of all resource entities, maintaining simplicity while enabling system-wide optimization.
Solution Approach 2:
Schedulers automatically collect and report their own resource consumption statistics to the coordinator, and use the distributed statistics to make their own scheduling decisions. This self-service approach enables global scheduling optimization without requiring external user agents to manually collect and manage statistics across the distributed system.
3Productivity
If resource weights are allocated to users and schedulers independently manage resources, then user quotas can be enforced, but performance requirements cannot be guaranteed under resource competition
Solution Approach 1:
The coordinator collects real-time resource consumption statistics from all schedulers and distributes this feedback information to relevant schedulers. This enables schedulers to adjust their resource allocation decisions dynamically, ensuring that user quotas are enforced while also guaranteeing performance requirements by preventing resource starvation under competitive conditions.
Solution Approach 2:
The system pre-allocates resource weights to users and maintains statistics on resource consumption patterns. Based on this preliminary setup and ongoing statistics collection, the coordinator can proactively make scheduling decisions that ensure both quota enforcement and performance guarantees before resource conflicts arise, rather than reacting after problems occur.
Data Source
AI summary
According to a method for scheduling a user request in a distributed resource system, an apparatus, and a system that are provided by embodiments of the present invention, in a Tn+1 period, an Sd acquires, from a coordinator Gk of a user z, a resource Cz(Tn) that is consumed by a user z request in a Tn period, and the Sd schedules, according to ϕz, Cz(Tn), Cz,d(Tn), and Nz,d(Tn), a Piz,d by using a scheduling algorithm. The user z request can be scheduled without depending on a user agent. In addition, the Sd schedules, according to ϕz, Cz(Tn), Cz,d(Tn), and Nz,d(Tn), the Piz,d by using the scheduling algorithm, thereby implementing global scheduling on the user z request and ensuring a performance requirement of the user z.