Hard Disk Drive Scheduling for Power and Latency Trade-offs

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional cold storage systems using hard disk drives or solid state drives face high costs and power consumption, and existing scheduling methods introduce significant switching times between disk groups, leading to increased processing delays and power inefficiencies.

Innovation Solution

A storage system with a controller, disk array, power supply, and heat dissipation system that schedules hard disk drives in a manner where only some groups are actively spinning while others are in a spin-up state, reducing power consumption and processing delays by allowing simultaneous active and spin-up states in different disk groups.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple disk groups are allowed to spin up simultaneously to reduce processing delays, then service processing speed is improved, but system power consumption increases

Engineering Contradiction:
Improveservice processing speedVSAvoidsystem power consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary actions by spinning up disk groups in advance before they are actually needed for processing. When service requests are received, the scheduler checks if the required disk groups are already in the active state. If not, it initiates a spin-up operation beforehand, allowing the disk group to be ready when needed, thus reducing actual processing delays while managing power consumption through advance preparation rather than simultaneous operation of all disk groups.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts the state of disk groups based on real-time service request patterns and system load. The scheduler continuously monitors which disk groups are active and which are inactive, and dynamically transitions disk groups between inactive and active states based on incoming service requests. This dynamic management allows the system to optimize the balance between processing speed and power consumption by having disk groups become active only when needed rather than remaining constantly active.

Inventive Principle:
Principle #15Dynamics

2Adaptability or versatility

If disk groups are switched frequently to process different service requests, then service request handling flexibility is improved, but processing delay increases due to switching time

Engineering Contradiction:
Improveservice request handling flexibilityVSAvoidprocessing delay
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The system performs preliminary actions by pre-spinning up disk groups that are likely to be needed based on service request patterns. Before a disk group is switched to process service requests, it is already in the active state, eliminating the switching delay that would otherwise occur. This preliminary preparation ensures that when service requests are directed to a disk group, it is immediately ready to process, thus maintaining flexibility while reducing processing delays.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system maintains continuity of useful action by keeping disk groups in the active state for extended periods rather than switching them on and off frequently. Once a disk group is spun up, it remains active and continues to process service requests without interruption or switching delays. This continuous operation eliminates the harmful switching time while maintaining the ability to handle different service requests, as multiple disk groups can be kept active simultaneously based on workload demands.

Inventive Principle:
Principle #20Continuity of useful action

3Use of energy by moving object

If only one disk group is allowed to work at a time to reduce power consumption, then system power is reduced, but service processing efficiency decreases

Engineering Contradiction:
Improvesystem powerVSAvoidservice processing efficiency
Core Design Contradiction:
Use of energy by moving objectVSProductivity

Solution Approach 1:

The system applies partial action by allowing a limited number of disk groups to be active simultaneously rather than all disk groups or just one. The scheduler controls the exact number of active disk groups based on service request workload, power consumption thresholds, and system capacity. This partial activation approach enables the system to process multiple service requests in parallel (improving efficiency) while constraining the total power consumption by limiting the number of simultaneously active disk groups to a manageable subset.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The system changes the operational parameters of disk groups dynamically based on service request patterns and power consumption levels. Instead of fixing the number of active disk groups at one value, the scheduler adjusts this parameter in real-time, increasing the number of active disk groups when service demand is high and decreasing it when demand is low or power consumption thresholds are approached. This parameter adjustment enables flexible balancing between power consumption and service processing efficiency.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3367244B1Storage system, storage device, and hard disk drive scheduling method
Publication Date: 2023.10.25 HUAWEI TECH CO LTD
  • EP3367244B1 patent drawingFigure 1
  • EP3367244B1 patent drawingFigure 2
  • EP3367244B1 patent drawingFigure 3

AI summary

This application discloses a storage system, a storage device, and a hard disk drive scheduling method. The storage system includes M disk groups, a power supply system, and a controller, and M is a natural number not less than 3. Each of the M disk groups includes at least one hard disk drive configured to store data, and each hard disk drive includes an inactive state, a spin up state, and an active state. Power supplied by the power supply system to the storage system allows one of the M disk groups to be in the active state and another disk group to be in the spin up state. The controller is connected to the M disk groups and configured to: control a hard disk drive in a first disk group of the M disk groups to be in the active state to process a first group of service requests, and when the hard disk drive in the first disk group is in the active state, control a hard disk drive in a second disk group of the M disk groups to switch from the inactive state to the spin up state. The second disk group is configured to process a second group of service requests.