Telemetry Data-Tolerance Tuning for Lower CPU and I/O Load

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Periodic data collection in information handling systems can negatively impact system performance due to increased CPU cycles and I/O access, particularly when multiple applications request data simultaneously, necessitating an improvement in the telemetry system's data collection mechanism.

Innovation Solution

A telemetry service adjusts data tolerance values based on criteria such as performance mode, number of active data collection policies, user presence, and system health to optimize data collection, reducing the frequency of data reads and minimizing performance impact.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If periodic data collection is performed to meet multiple application requests, then data availability and service capability are improved, but system performance deteriorates due to increased CPU cycles and I/O access

Engineering Contradiction:

Improvedata collection capabilityVSAvoidsystem performance

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent combines multiple data collection requests into a single unified request when multiple applications need the same data. The telemetry service consolidates duplicate requests and performs one data read operation to satisfy multiple applications, thereby reducing CPU cycles and I/O access while maintaining data availability for all requesting applications.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The system performs preliminary data collection and caching before applications actually need the data. By proactively collecting data based on predicted needs or scheduled intervals, the system prepares data in advance, reducing the need for frequent real-time data reads and minimizing performance impact during actual data retrieval operations.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If data collection frequency is increased to improve data freshness, then data quality is improved, but system load increases due to more CPU cycles and I/O access

Engineering Contradiction:

Improvedata freshnessVSAvoidsystem load

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent applies partial data collection by selecting only the specific data points needed by applications rather than collecting all available data. The telemetry service analyzes application requirements and collects only the necessary subset of data, reducing overall system load while ensuring that each application receives the specific data it needs with appropriate freshness.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

Different data types and sources are collected at different frequencies based on their specific requirements. Critical data that needs to be fresh is collected more frequently, while less critical data is collected less often. This localized optimization of collection frequency reduces total system load while maintaining data freshness where it matters most.

Inventive Principle:

Principle #3Local quality

Data Source

PatentUS20250292264A1Optimizing data tolerance value

Publication Date: 2025.09.18 DELL PROD LP

AI summary

An information handling system generates a data tolerance policy with a data tolerance value for a data type. If a data collection request meets a set of criteria that includes whether a number of active data collection policies exceeds a threshold, then the system modifies the data tolerance value of the data type in the data tolerance policy. The system instruments data according to a modified data tolerance value, the data collection request, and the data tolerance policy.