Distributed Storage Disk Clusters Energy Saving States
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current storage disk systems for warm data face challenges in balancing low power consumption with low data latency, as they cannot be powered off or set to a dormant state due to millisecond-level access latency requirements, leading to high power consumption and long wake-up times.
Innovation Solution
A distributed storage system with multiple storage disk clusters configured in different energy-saving states, where data is split into sub-data and stored across clusters with varying wake-up latencies to minimize power consumption while meeting latency requirements by sequential reading based on access sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the storage disk is kept in a readable state to meet millisecond-level access latency requirements, then data access speed is improved, but power consumption increases
Solution Approach 1:
The patent divides data into hot data and warm data, and further segments warm data into different sub-data portions. Different segments are stored in storage disks with different energy-saving states configured, allowing the system to meet latency requirements for frequently accessed portions while reducing power consumption for less frequently accessed portions.
Solution Approach 2:
The patent applies different energy-saving state configurations to different storage disk clusters based on their specific roles and access patterns. Storage disks storing sub-data with higher access priority are configured with shorter wake-up latencies, while others use longer wake-up latencies, optimizing the balance between performance and power consumption locally.
2Use of energy by moving object
If the storage disk is powered off or set to dormant state to reduce power consumption, then energy savings are achieved, but wake-up time increases to 30-50 seconds
Solution Approach 1:
The patent dynamically configures different energy-saving states with varying wake-up latency characteristics for different storage disk clusters. This allows the system to adapt between power-saving modes and performance modes based on data access requirements, avoiding the extreme of complete power-off while still achieving significant power reduction.
Solution Approach 2:
The system proactively manages the energy states of storage disks based on predicted access patterns and data importance. By pre-configuring appropriate energy-saving states for different data segments, the system avoids last-minute wake-up operations that would cause latency issues.
3Speed
If all storage disks are kept in readable state, then data access latency is minimized, but overall system power consumption is high
Solution Approach 1:
The patent segments the storage system into multiple storage disk clusters, each managing different portions of warm data with different access priorities. This segmentation allows the system to optimize power consumption at the cluster level rather than maintaining all disks in high-performance state, reducing overall system power consumption while maintaining acceptable access latency for critical data.
Solution Approach 2:
The patent changes the energy-saving state parameters (specifically wake-up latency) of different storage disk clusters based on the access patterns and importance of the data they store. This parameter optimization allows the system to achieve an optimal balance between aggregate performance and power consumption.
Data Source
AI summary
A data processing method is provided. The method is implemented in a distributed storage system including at least two storage disk clusters. The at least two storage disk clusters are configured with at least two energy saving states. The method includes: receiving, by a processing module of the distributed storage system, a read request including an identifier of to-be-processed data; reading metadata of the to-be-processed data based on the identifier of the to-be-processed data, to determine a first storage disk cluster and a second storage disk cluster, where the first storage disk cluster and the second storage disk cluster are configured with different energy saving states; and reading first sub-data from the first storage disk cluster, and after the first sub-data is read, reading second sub-data from the second storage disk cluster.