Non-Intrusive Idle Processor Detection via Power Consumption ML

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing solutions for identifying idle computing devices in data centers are intrusive, causing system overhead and resource consumption, and can be terminated by administrators, as they require agents to run within the operating system to collect metric information.

Innovation Solution

A non-intrusive approach using machine learning to analyze power consumption data from service processors, such as baseboard management controllers, to classify computing devices as idle or busy, leveraging Intelligent Platform Management Interface technology for data collection outside the operating system.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If an agent or process runs within the operating system to collect metric information, then the system can obtain detailed usage metrics, but system overhead increases and the solution can be terminated by administrators

Engineering Contradiction:

Improveusage metrics collectionVSAvoidsystem overhead

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces a network device as an intermediary that collects power consumption data from computing devices. Instead of running agents within the operating system, the network device receives power consumption information directly from the hardware, acting as a mediator between the computing devices and the analysis system. This eliminates the need for intrusive software agents while maintaining data collection capabilities.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the software-based agent system with a hardware-based power consumption monitoring approach. By substituting the mechanical/software system (agents running in OS) with a physical measurement system (power consumption data collection at hardware level), the solution eliminates system overhead and administrative termination risks while maintaining measurement precision.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If an agent or process runs within the operating system to collect metric information, then the system can obtain detailed usage metrics, but the solution consumes bandwidth and resources on the primary network connection

Engineering Contradiction:

Improveusage metrics collectionVSAvoidnetwork bandwidth consumption

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The network device serves as an intermediary that collects power consumption data directly from computing devices through dedicated hardware monitoring capabilities. This intermediary approach allows data collection without utilizing the primary network connection for metric transmission, thereby conserving network bandwidth and resources while maintaining precise measurement capabilities.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the network-based metric collection mechanism with a direct hardware monitoring approach. By substituting the software agent that transmits data over the network with a hardware-based power consumption monitoring system, the solution eliminates network bandwidth consumption while maintaining measurement precision through direct physical measurement.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If an agent or process runs within the operating system, then the system can collect metric information, but the agent can be terminated or stopped by administrative users

Engineering Contradiction:

Improvedata collection continuityVSAvoidadministrative control

Core Design Contradiction:

ReliabilityVSEase of operation

Solution Approach 1:

The patent replaces the software-based agent system with a hardware-based power consumption monitoring system. By substituting the terminable software agent with a hardware monitoring mechanism that operates at the physical level, the solution ensures continuous and reliable data collection that cannot be terminated by administrative users, while maintaining ease of operation through centralized control interfaces.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS20230342618A1Identifying idle processors using non-intrusive techniques

Publication Date: 2023.10.26 NVIDIA CORP

AI summary

Apparatuses, systems, and techniques to classify computing devices as idle or busy using a machine learning (ML) model based on power consumption data collected non-intrusively are described. One method receives power consumption data for a computing device from a service processor operatively coupled to the computing device. The method determines a set of features from the power consumption data for a first time period. The method classifies, using an ML model and the set of features, whether the computing device is idle or busy in the first time period. The method outputs an indication of the computing device being idle responsive to a classification that the computing device is idle.