Request Scheduler Deadline Management for Storage Bandwidth
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage resource arrangements face challenges in guaranteeing bandwidth for bandwidth-intensive applications like streaming video, as the non-deterministic nature of storage access makes it difficult to provide a minimum guaranteed bandwidth without over-provisioning, leading to inefficient resource utilization and unsatisfactory user experiences.
Innovation Solution
A request scheduler system that manages client requests in a multi-device storage system by setting deadlines based on service IDs and payload sizes, sorting requests into FIFO queues, and advancing selected requests for immediate processing when deadlines are met, ensuring aggregated minimum bandwidth guarantees across distributed service stations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If storage resources are heavily over-provisioned to guarantee minimum bandwidth, then bandwidth reliability is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The system performs preliminary actions by pre-calculating deadlines for storage requests based on guaranteed bandwidth requirements before requests are submitted. The request scheduler computes deadline values and stores them in advance, allowing the system to meet bandwidth guarantees without needing excessive spare capacity. This proactive deadline assignment enables efficient resource allocation while maintaining reliability.
Solution Approach 2:
The system implements dynamic scheduling by continuously monitoring request deadlines and adjusting the processing order of storage requests in real-time. The request scheduler dynamically selects which requests to process next based on their deadline values, ensuring that bandwidth-intensive applications receive appropriate service while optimizing overall resource utilization. This dynamic approach replaces static over-provisioning with adaptive resource management.
2Loss of energy
If storage resources are not over-provisioned, then resource utilization efficiency is improved, but bandwidth reliability deteriorates
Solution Approach 1:
The system changes the parameter of request scheduling from traditional FIFO or priority-based methods to deadline-based scheduling. By introducing deadline values calculated from guaranteed bandwidth requirements, the system transforms how storage requests are processed. This parameter change enables the system to provide bandwidth guarantees with normal resource provisioning rather than requiring over-provisioning, thus maintaining both efficiency and reliability.
3Device complexity
If traditional FIFO scheduling is used, then system simplicity is maintained, but bandwidth guarantee capability deteriorates
Solution Approach 1:
The system segments the storage request processing by dividing requests into different categories based on their deadline values. The request scheduler processes requests in segments according to their urgency, with deadline-based prioritization ensuring that bandwidth-intensive requests receive appropriate attention. This segmentation approach maintains relative system simplicity while enabling bandwidth guarantees, as it builds upon the familiar FIFO concept but adds deadline-based segmentation for critical requests.
Data Source
AI summary
Apparatus and method for scheduling requests for data transfers in a multi-device storage system. In some embodiments, a system includes at least one server coupled to a pool of storage devices to transfer data from the storage devices to client devices responsive to requests. A request scheduler is adapted to receive into a memory a plurality of requests each having a service identifier (ID) and a payload size, to set a deadline for each request responsive to the service ID and the payload size, to forward the requests to the server for processing in an order based on service ID and, responsive to the deadline being reached for a selected request, to advance the selected request for immediate processing by the server.