Memory Controller Workload Scheduling for User Fairness
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory controllers struggle to ensure fair sharing of storage devices among multiple users, as they lack accurate workload determination and scheduling mechanisms to prioritize commands based on resource consumption.
Innovation Solution
A memory controller that includes a command-unit workload calculator to determine workloads by unit of each command, a user-unit workload calculator to calculate workloads for each user, and a command fetch unit that prioritizes commands from users with relatively small workloads, ensuring fair sharing of storage devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a storage device is shared by multiple users without workload-based scheduling, then device utilization is high, but user fairness deteriorates
Solution Approach 1:
The patent segments the storage device workload into individual user workloads by introducing a user-unit workload calculator that divides total workload into separate components for each user. This segmentation enables independent tracking and fair scheduling of each user's resource consumption, resolving the fairness issue while maintaining overall high utilization.
Solution Approach 2:
The patent implements dynamic scheduling by continuously calculating workloads and adjusting command fetch priorities in real-time. The command fetch unit dynamically changes fetch order based on current workload states, allowing the system to adapt to varying user demands and maintain fairness without reducing overall device utilization.
2Device complexity
If commands are fetched in FIFO order without workload consideration, then scheduling is simple, but resource consumption fairness deteriorates
Solution Approach 1:
The patent introduces feedback mechanisms where the workload calculator continuously monitors resource consumption and feeds this information back to the command fetch unit. This feedback loop enables fair resource allocation by adjusting fetch priorities based on actual workload measurements, achieving fairness without excessive complexity through automated closed-loop control.
Solution Approach 2:
The patent changes the scheduling parameter from simple FIFO order to workload-based priority ordering. By calculating workload metrics and using these as the basis for command fetch sequencing, the system achieves fair resource consumption distribution while maintaining manageable complexity through systematic parameter transformation.
3Reliability
If workload-based priority scheduling is implemented, then user fairness is improved, but calculation overhead increases
Solution Approach 1:
The patent applies partial action by calculating workload only for the specific memory operations actually performed, rather than evaluating all possible operations. The workload calculator focuses on measuring actual resource consumption from executed commands, reducing unnecessary calculation overhead while maintaining accurate fairness assessment.
Data Source
AI summary
A memory controller with improved user fairness, a storage device including the memory controller, and an operating method of the memory controller are provided. A memory controller includes a command-unit workload calculator configured to calculate a workload by a unit of each command, based on resource consumption of a memory device, which is determined by performing at least one memory operation included in each of the commands, a user-unit workload calculator configured to calculate a workload caused by each of a plurality of users, based on the workload calculated by the unit of the command, and a command fetch unit configured to preferentially fetch, from a host, a command issued by a user who has caused a relatively small workload, based on calculation results of workloads of the plurality of users.