Distributed Storage Power Reduction via Dynamic Replica Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data centers face high power consumption due to non-power proportional storage systems, which limits the overall power proportionality of the data center, despite efforts like dynamic voltage scaling for server CPUs.
Innovation Solution
A storage system that reduces power consumption by dividing data into chunks, replicating them across servers, and reducing the number of active replicas during low load periods, with servers in a low power state, and using a versioned store for writes to be copied back when servers return to normal power state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If the number of active replicas is reduced during low load periods, then power consumption is reduced, but data availability and access speed may be affected
Solution Approach 1:
The system dynamically adjusts the number of active replicas based on workload conditions. During low load periods, replicas are deactivated to reduce power consumption, while during high load periods, all replicas are activated to ensure data availability and performance. This dynamic adaptation resolves the contradiction by making the system flexible rather than static.
Solution Approach 2:
The system performs preliminary actions by pre-loading data into a versioned store before replicas are deactivated. This ensures that when replicas are inactive, the versioned store already contains the necessary data to maintain availability. The preliminary preparation of data in the versioned store prevents service disruption when replicas are reduced.
2Use of energy by stationary object
If servers are put into low power state, then power consumption is reduced, but write operation latency increases
Solution Approach 1:
The versioned store acts as an intermediary between clients and inactive replicas. When a server is in low power state, write operations are redirected to the versioned store instead of the inactive replica. This intermediary mechanism allows the system to maintain write functionality while servers remain powered down, eliminating the latency penalty that would otherwise result from waking servers for each write operation.
Solution Approach 2:
The system creates copies of data in the versioned store as a backup mechanism. When replicas are deactivated, the versioned store holds copies of the data that can be used to service write operations. This copying approach allows the original replicas to remain inactive while maintaining system functionality.
3Use of energy by stationary object
If the number of active replicas is reduced, then power consumption is reduced, but system complexity increases due to versioned store management
Solution Approach 1:
The versioned store serves multiple functions: it acts as a backup storage during replica deactivation, a data synchronization point when replicas are reactivated, and a mechanism for maintaining data consistency across the distributed system. This multi-functionality reduces the need for separate specialized components, thereby limiting the increase in system complexity despite the added capability.
Data Source
AI summary
Methods for reducing the power consumption of distributed storage systems are described. An embodiment describes a storage system which is adapted to reduce its power consumption at times of low load by reducing the number of active versions of the stored data. The data to be stored in the storage system is divided into chunks and in an example, each chunk is replicated on a number of different servers. At times of low load, the system enters a mode of operation in which the number of active replicas is reduced and servers that do not store any active replicas are put into a low power state. When in this mode, writes are written to a versioned store and the data is subsequently copied to servers storing replicas once all the servers have returned to normal power state.