Storage Write Off-Loading for Energy Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face high power consumption due to storage systems, with storage sub-systems lacking advanced power management, leading to inefficiencies as disks remain powered and consume energy even when idle, exceeding processor power consumption.
Innovation Solution
Implementing a write off-loading method where data is temporarily stored in alternative locations when the intended storage location is unavailable, allowing disks to be spun down during idle periods, reducing power usage by redirecting writes to other storage within the system and reclaiming data when the original location becomes available.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If disks are kept powered on to ensure immediate write availability, then write response time is improved, but power consumption increases significantly
Solution Approach 1:
The system performs preliminary actions by maintaining a pool of standby storage devices that are powered on and ready to accept writes immediately. When the primary storage is full or slow, writes are redirected to these pre-prepared standby devices without waiting for power-on sequences, thus maintaining fast response times while allowing the primary storage to be powered down during idle periods
Solution Approach 2:
The system introduces a metadata service and distributed file system layer as intermediaries between clients and storage devices. This abstraction layer tracks the location and availability of data across multiple devices, enabling seamless redirection of writes to alternative storage locations when primary devices are unavailable or powered down, without impacting client performance
2Use of energy by moving object
If disks are spun down during idle periods to reduce power consumption, then energy efficiency is improved, but write availability deteriorates when storage is needed
Solution Approach 1:
The system merges multiple storage devices into a unified distributed storage pool where data can be written to any available device. By combining the capacities of multiple devices and using a distributed file system, the system ensures that even if some devices are powered down for energy savings, other devices in the pool remain available to accept writes, thus maintaining write availability while improving energy efficiency
Solution Approach 2:
The system dynamically changes the operational parameters of storage devices based on workload conditions. During idle periods, devices are transitioned to low-power states to improve energy efficiency. When write operations are needed, the system detects the demand and activates appropriate devices, changing their power state parameter from idle to active based on real-time storage pool conditions
3Reliability
If data is redirected to alternative storage locations when primary storage is unavailable, then storage availability is maintained, but system complexity increases
Solution Approach 1:
The system implements universal storage devices that can serve multiple functions: primary storage, standby storage, and temporary write targets. Each storage device in the distributed pool is designed to be multi-functional, capable of accepting writes directly or receiving redirected writes from failed devices. This universality simplifies the redirection logic compared to having specialized dedicated backup devices for each primary storage unit
Data Source
AI summary
Improved storage systems which use write off-loading are described. When a request to store some data in a particular storage location is received, if the particular storage location is unavailable, the data is stored in an alternative location. In an embodiment, the particular storage location may be unavailable because it is powered down or because it is overloaded. The data stored in the alternative location may be subsequently recovered and written to the particular storage location once it becomes available.