Rack Manager Power Throttling for Datacenter Failover
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Datacenters face inefficiencies due to the need for separate racks for different applications with varying power utilization levels, leading to increased costs and complexity in managing power distribution during power feed failures.
Innovation Solution
A unified server and rack architecture with a rack manager that monitors and adjusts power consumption by throttling or shutting off servers to maintain power within safe thresholds, preventing circuit breaker tripping during power feed failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If servers are configured to consume high power to meet application demands, then application performance is improved, but the risk of power overdraft during power feed failure increases
Solution Approach 1:
The system performs preliminary actions by pre-identifying which servers should be throttled or shut off before actual power failure occurs. The rack manager maintains a list of candidate servers that can be deactivated, and when power feed failure is detected, the system already has a prepared plan to reduce power consumption by deactivating specific servers, thereby preventing power overdraft while maintaining application performance.
Solution Approach 2:
The system implements feedback mechanisms where the rack manager continuously monitors power consumption levels and compares them against thresholds. When power feed failure is detected or power consumption approaches critical levels, the system receives feedback and automatically adjusts server operation by throttling or shutting off specific servers to bring power consumption back within safe limits.
2Use of energy by moving object
If separate racks are used for different applications with varying power utilization, then power distribution is optimized, but datacenter complexity and costs increase
Solution Approach 1:
The system implements universality by enabling a single rack to host multiple applications with different power utilization characteristics. Instead of requiring separate racks for different applications, the rack manager dynamically allocates power feeds to servers based on real-time demands, allowing one rack to serve multiple application types efficiently while maintaining optimized power distribution.
Solution Approach 2:
The system introduces dynamics by making power allocation flexible and adaptive rather than static. The rack manager continuously adjusts which servers are active, throttled, or shut off based on current power feed availability and application demands, allowing the same physical rack configuration to adapt to varying power requirements of different applications without physical reconfiguration.
3Object-affected harmful factors
If circuit breakers are installed to prevent power overdraft, then power safety is improved, but the risk of complete power loss during failover increases
Solution Approach 1:
The rack manager acts as an intermediary between the power feeds and the servers. Instead of relying solely on circuit breakers to protect against power overload, the rack manager proactively monitors power consumption and mediates power distribution by throttling or shutting off specific servers before the circuit breaker would trip, thereby preventing complete power loss while maintaining protection against overload.
Solution Approach 2:
The system takes preliminary action by identifying and deactivating specific servers before power feed failure causes circuit breaker tripping. The rack manager pre-determines which servers should be throttled or shut off to bring power consumption below the circuit breaker threshold, thereby preventing the harmful effect of complete power loss while maintaining overload protection.
Data Source
AI summary
A rack in a datacenter is powered by a first power feed and a second power feed. The rack supports a plurality of servers which have a maximum combined power consumption which is greater than a maximum supplied power from either the first power feed or the second power feed. When power is lost from one of the power feeds, a rack manager reduces the total power consumption of the plurality of servers by throttling at least one of the servers and/or shutting off at least one of the plurality of servers.