Storage Cluster Power Conservation via Node Bidding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage clusters, it is challenging to conserve power while maintaining access for client applications and ensuring digital object integrity, especially during periods of low usage when existing methods either compromise throughput or require complex administrative arrangements.
Innovation Solution
Implementing a power conservation technique that allows nodes and volumes in a storage cluster to independently decide when to spin down disks and reduce CPU utilization based on usage patterns, using mechanisms like sleepAfter and wakeAfter parameters, and bidding processes to manage resource allocation, thereby reducing power consumption without affecting maximum throughput.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If the storage cluster keeps all nodes and disk drives active to maintain maximum throughput, then client application access and object integrity are ensured, but power consumption remains high
Solution Approach 1:
The patent implements dynamic power management by allowing nodes and volumes to transition between active and idle states based on usage patterns. Nodes monitor their own utilization and automatically spin down disk drives and reduce CPU utilization during idle periods, then wake up when needed, creating a dynamic system that adapts to workload changes rather than maintaining a static active state
Solution Approach 2:
Each node and volume makes autonomous decisions about when to enter idle state and when to wake up based on its own utilization metrics and configured parameters (sleepAfter, wakeAfter). The system does not require external management or complex administrative arrangements to coordinate power savings, as each component self-manages its power state based on local conditions
2Use of energy by moving object
If the storage cluster spins down disk drives to conserve power during low usage, then power consumption is reduced, but access latency increases when nodes need to wake up
Solution Approach 1:
Nodes perform preliminary actions by monitoring their own utilization patterns and proactively entering idle state before complete inactivity occurs. The sleepAfter parameter allows nodes to anticipate low-usage periods and prepare for power savings, while still maintaining the ability to quickly wake up when needed, thus reducing the impact of spin-up latency
Solution Approach 2:
The patent allows administrators to configure parameters (sleepAfter, wakeAfter, idleCost) that control the timing and conditions of power state transitions. By adjusting these parameters, the system can optimize the balance between power savings and access latency based on specific workload patterns and performance requirements
3Ease of operation
If the storage cluster allows decentralized autonomous decisions by nodes about when to sleep and wake, then administrative overhead is reduced, but system complexity increases
Solution Approach 1:
Each node and volume autonomously monitors its own utilization and makes independent decisions about when to enter idle state and when to wake up, based on configured parameters and local conditions. This self-service approach eliminates the need for complex centralized coordination or administrative intervention, as each component manages its own power state
Solution Approach 2:
The patent divides the storage cluster into independent nodes and volumes that each make autonomous decisions about their power state. This segmentation allows each component to operate independently based on its own utilization patterns, simplifying the overall system management while maintaining coordinated behavior through standardized protocols
Data Source
AI summary
Each node and volume in a storage cluster makes a decision whether to reduce power consumption based on lack of requests from client applications and nodes over a time period. Node configuration parameters determine how long to wait until idling a node or volume, and how long to wait while idle before performing integrity checks. A bid value is calculated by each node and reflects how much it will cost for that node to write a file, read a file, or keep a copy. A node with the lowest bid wins, and nodes that are idle have a premium added to each bid to ensure that idle nodes are kept idle. In an archive mode, writes bids are reversed, nodes with less capacity submit lower bids, fuller nodes fill up faster and are then idled, while empty or near empty nodes may remain idle before winning a write bid.