Data Center Power Management via Dynamic Clock and Voltage Scaling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data centers consume high levels of power due to over-provisioning for peak demand, as existing load management features are not responsive enough to real-time changes in application workload, leading to inefficient power usage during periods of low demand.
Innovation Solution
A power and capacity management utility dynamically adjusts the clock rate and processor supply voltage of servers, and activates or deactivates them based on performance characteristics and predicted demand, to maintain performance within predefined targets while minimizing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If servers are provisioned with sufficient hardware to meet maximum demand, then service level agreements are satisfied, but power consumption increases during low demand periods
Solution Approach 1:
The system dynamically adjusts server clock rates and activates/deactivates processors based on real-time workload monitoring. When demand is low, clock rates are reduced and processors are deactivated; when demand increases, clock rates are increased and processors are activated, allowing the system to adapt its power consumption to actual needs while maintaining SLA compliance
Solution Approach 2:
The invention changes operational parameters (clock rate, processor activation state) of servers based on measured workload characteristics. By monitoring transaction volumes and response times, the system adjusts these parameters to maintain performance within acceptable ranges while optimizing power usage
2Stability of the object's composition
If conventional load management features are used, then system stability is maintained, but responsiveness to real-time workload changes is insufficient
Solution Approach 1:
The system implements continuous feedback loops that monitor workload characteristics (transaction volumes, response times) and use this information to adjust server clock rates and processor activation states. This closed-loop control enables rapid response to workload changes while maintaining system stability through controlled adjustments
Solution Approach 2:
The system proactively adjusts server parameters before performance degradation occurs by monitoring workload trends and predicting future demands. This allows the system to prepare appropriate clock rate adjustments and processor activations in advance, maintaining stability while responding quickly to anticipated changes
Data Source
AI summary
A method of managing data center power usage based on service commitments. A power and capacity management utility measures performance characteristics of applications hosted by a data center that include power consumption and transaction processing time. When the performance characteristics are above (or below) a first pre-defined target range, the utility decreases (or increases, respectively) a clock rate and processor voltage of one or more active servers. When the performance characteristics are within the first target range and the clock speed is above (or below) a second pre-defined target range, the utility activates (or deactivates, respectively) one or more inactive (or active, respectively) servers. When the size of an inactive server pool is above a third pre-defined target range, the utility moves inactive servers to an unallocated server pool. When the size of the inactive server pool is below the third target range, the utility allocates additional servers.