Supercapacitor Power Spike Management for Multi-Node Compute Systems

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Multi-node compute systems face challenges in efficiently managing power supply, as existing designs often require oversized power infrastructure to accommodate peak power demands, leading to increased costs and potential system crashes during rare concurrent peak events.

Innovation Solution

A shared power supply system combined with an auxiliary power source, such as a supercapacitor or battery, is used to manage power spikes, allowing the power supply to be sized closer to actual consumption levels while ensuring system survivability during peak events by providing a well-defined and specified power spike support.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If a shared power supply is sized for the sum of maximum peak power (Pmax) of all compute nodes, then system reliability is improved (preventing system crashes during concurrent Pmax events), but system cost and power budget increase undesirably

Engineering Contradiction:

Improvesystem reliabilityVSAvoidsystem cost

Core Design Contradiction:

ReliabilityVSQuantity of substance

Solution Approach 1:

The power supply system is segmented into two distinct components: a shared power supply that handles baseline power consumption, and individual uninterruptible power supplies (UPS) at each compute node that handle peak power events. This segmentation allows the shared power supply to be sized for average consumption rather than peak consumption, reducing overall system cost while maintaining reliability during concurrent peak events.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Each compute node is equipped with a local UPS that is pre-charged and ready to immediately supply power during peak events. This preliminary action ensures that when a compute node experiences a Pmax event, the local UPS can instantly bridge the power gap without requiring the shared power supply to be oversized, thus maintaining system reliability without increasing system cost.

Inventive Principle:

Principle #10Preliminary action

2Reliability

Engineering Contradiction:

Improvesystem reliabilityVSAvoidpower consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The power supply system is segmented into a shared power supply for baseline consumption and local UPS units for peak consumption. This allows the shared power supply to be sized according to actual average consumption levels rather than theoretical maximum consumption, thereby reducing unnecessary power consumption while maintaining system reliability during peak events.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Each compute node serves its own peak power needs through a local UPS unit, rather than relying on the shared power supply to provide excessive capacity. This self-service approach allows each node to independently handle its own power spikes, reducing the overall power consumption of the system while maintaining reliability.

Inventive Principle:

Principle #25Self-service

3Reliability

If an auxiliary power source is added to provide power spike support, then system survivability during peak events is improved, but device complexity increases

Engineering Contradiction:

Improvesystem survivabilityVSAvoiddevice complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The complexity of power management is segmented and distributed to individual compute nodes through local UPS units, rather than requiring a complex centralized power supply system. Each node independently manages its own peak power events, simplifying the overall system architecture while improving survivability during peak events.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Local UPS units act as intermediaries between the shared power supply and compute nodes during peak events. These intermediaries absorb the complexity of peak power management, allowing the shared power supply to remain simple while ensuring system survivability during concurrent Pmax events.

Inventive Principle:

Principle #24Intermediary (Mediator)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach reduces system costs and power consumption while maintaining reliability, allowing for smaller, more efficient power supplies that can handle power spikes without causing system crashes, thereby enhancing energy efficiency and reducing the risk of electrical failures.

Implementation Method 1

An auxiliary power source, such as a supercapacitor or battery, is used to manage power spikes

Methodology Applied

Scientific EffectCapacitance: Capacitance

Data Source

PatentUS10761579B2Supercapacitor-based power supply protection for multi-node systems

Publication Date: 2020.09.01 INTEL CORP

AI summary

In one embodiment, a system includes: a plurality of compute nodes to couple in a chassis; a first shared power supply to provide a baseline power level to the plurality of compute nodes; and an auxiliary power source to provide power to one or more of the plurality of compute nodes during operation at a higher power level than the baseline power level. Other embodiments are described and claimed.