Virtual Machine Latency Decomposition via Pre-Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to effectively decompose aggregate latency measurements into their additive components (host latency, kernel latency, network latency, and storage latency) for virtual machines, making it difficult for system administrators to diagnose the primary cause of latency issues in storage platforms.
Innovation Solution
A method to preprocess sensor measurements, including data interpolation, averaging, and time shifting, to isolate and display these latency components, allowing for the decomposition of aggregate latency into host latency, kernel latency, network latency, and storage latency, facilitating visual diagnosis of latency causes over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If aggregate latency measurements are used without decomposition, then the measurement process is simple, but the diagnostic capability is insufficient
Solution Approach 1:
The patent segments the aggregate latency measurement into distinct additive components: host latency, kernel latency, network latency, and storage latency. Each component is measured and reported separately by different system elements (virtual disk, datastore, storage system), allowing precise identification of latency sources without overwhelming complexity through structured decomposition
Solution Approach 2:
The patent introduces intermediary processing mechanisms including pre-processing modules that perform data interpolation, averaging, and time-shifting, as well as supplemental information generation that correlates latency components with system events. These intermediaries transform raw measurements into diagnostically useful information while managing complexity through standardized processing pipelines
2Loss of information
If latency components are decomposed and pre-processed, then the diagnostic capability is improved, but the data processing complexity increases
Solution Approach 1:
The patent applies preliminary actions through pre-processing operations performed on latency measurements before they are combined and analyzed. Data interpolation fills missing values, averaging smooths variations, and time-shifting aligns measurements from different sources. These preliminary actions ensure information completeness and consistency while organizing complexity into standardized preprocessing routines
Solution Approach 2:
The patent implements feedback mechanisms where processed latency information is displayed with supplemental details that provide context for diagnosis. The system continuously monitors and adjusts based on the correlation between latency components and system events, creating a feedback loop that improves diagnostic accuracy while managing processing complexity through iterative refinement
Data Source
AI summary
Described herein are techniques to decompose a measurement of the latency experienced by a virtual machine into its additive components (e.g., host latency, kernel latency, network latency and storage latency). The host latency may be determined as the aggregate latency reported from the virtual disk of the virtual machine minus the aggregate latency reported from the datastore on which the virtual disk is stored. The network latency may be determined as the device latency reported from the datastore minus the storage latency reported by a storage system. Various sensor measurements (e.g., aggregate latency reported from the virtual disk and/or datastore) may be first pre-processed to compensate for non-idealities before being combined with other sensor measurements.