Cloud Performance Metrics Aggregation via Topology Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing systems face challenges in aggregating performance metrics from individual components into logical components, leading to disjointed performance data that does not reflect the performance of services created from physical resources, and requiring metrics to be defined based on hardware components for effective resource usage insights.
Innovation Solution
A method and system for collecting performance data from cloud computing components, including fabric interconnects, switches, and storage arrays, transforming raw data into performance metrics using topology information, and creating logical layer metrics based on physical layer relationships, such as storage total throughput and utilization metrics for services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If performance metrics are collected from individual physical components, then detailed raw data is obtained, but the data becomes disjointed and does not reflect service-level performance
Solution Approach 1:
The patent introduces topology information as an intermediary that maps relationships between physical components and logical services. This intermediary enables the system to correlate raw component metrics with service-level performance by providing the contextual relationships needed to aggregate and interpret the data meaningfully.
Solution Approach 2:
The patent adds a new dimension of analysis by introducing topology information that connects physical components to logical services across multiple layers. This dimensional transformation allows performance data to be viewed and aggregated from both physical component perspectives and logical service perspectives simultaneously.
2Reliability
If raw performance data from all components is stored, then complete data availability is achieved, but data storage requirements increase significantly
Solution Approach 1:
The patent extracts only the essential topology information needed to map component relationships, rather than storing all possible raw performance data. By extracting and storing topology metadata that defines relationships, the system can derive service-level metrics without retaining complete raw data from every component.
Solution Approach 2:
The patent creates a topological model or copy of the system's structural relationships, which serves as a lightweight representation that enables performance aggregation without requiring storage of all original raw performance data from individual components.
3Measurement precision
If metrics are defined based on hardware components, then component-level monitoring is achieved, but service-level resource usage insights are lost
Solution Approach 1:
Topology information serves as the intermediary that bridges component-level metrics and service-level performance. It provides the mapping relationships that allow the system to translate hardware component data into meaningful service-level insights without losing the connection to underlying component performance.
Solution Approach 2:
The patent merges component-level performance data with topology relationship information to produce aggregated service-level metrics. This merging process combines multiple component metrics according to their topological relationships, creating unified service-level views while preserving the underlying component data integrity.
Data Source
AI summary
Methods and apparatus to provide performance data transformation in a cloud computing system. In one embodiment, the system performs data transformation with information from a configuration subsystem, to generate metrics for network layer, storage layer, compute layer, and logical components.