Resource management controller
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional scheduler technologies lack visibility into the actual utilization of hardware resources such as SSDs and HW accelerators, leading to non-optimal resource allocation and waste in hyper-scaler environments, with challenges like underutilization or overloading due to the use of virtual time-based scheduling.
Innovation Solution
Implementing a Resource Management Controller (RMC) that tracks actual resource utilization through run-time telemetry, allowing for priority-based and on-the-fly decisions to optimize resource allocation across multiple layers of the stack, including CPU, memory, and peripherals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If virtual time-based scheduling is used, then task scheduling is simplified, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements a resource management controller that collects run-time telemetry data from hardware resources and feeds this information back to the scheduler. This feedback mechanism enables the scheduler to adjust task allocation based on actual resource utilization, resolving the contradiction between scheduling simplicity and resource utilization efficiency.
Solution Approach 2:
The patent introduces a resource management controller as an intermediary between the scheduler and hardware resources. This mediator collects utilization data from resources and provides it to the scheduler, allowing the scheduler to make informed decisions without increasing operational complexity.
2Device complexity
If conventional scheduler technology is used, then system complexity is reduced, but resource allocation optimality deteriorates
Solution Approach 1:
The resource management controller serves as an intermediary layer between the conventional scheduler and hardware resources. It maintains the simplicity of the existing scheduler while introducing intelligence for optimal resource allocation through run-time data collection and analysis.
Solution Approach 2:
The patent segments the resource management function into a separate controller module that handles telemetry data collection and analysis. This segmentation allows the main scheduler to remain simple while the dedicated controller handles the complex task of optimizing resource allocation based on actual utilization.
3Ease of manufacture
If virtual time-based scheduling is used, then scheduling implementation is easier, but resource visibility deteriorates
Solution Approach 1:
The patent implements a feedback loop where the resource management controller continuously collects run-time telemetry data from hardware resources and makes this utilization information visible to the scheduling system. This resolves the information loss while maintaining the ease of virtual time-based scheduling implementation.
Data Source
AI summary
An embodiment of an electronic apparatus may comprise one or more substrates, and a controller coupled to the one or more substrates, the controller including circuitry to read utilization-related information for a resource from a memory shared with a processor in response to a request from the processor for the resource, and schedule utilization of the resource based at least in part on the utilization-related information for the resource. Other embodiments are disclosed and claimed.