Virtual Resource Metric Management in NFVI
Solution Overview
Problem
Current network function virtualization infrastructure (NFVI) lacks efficient mechanisms for collecting and managing virtual resource metrics across virtual machine instances, leading to suboptimal resource utilization and management in cellular communication networks.
Innovation Solution
Implementing a system where an agent on a compute node collects virtual resource data from the hypervisor and delivers it to the NFVI for conversion into usage metrics, which are then provided to a virtualized infrastructure manager (VIM) for transmission to a virtualized network function manager (VNFM), utilizing OpenStack Telemetry services to collect CPU, storage, and networking metrics through Cumulative and Gauge meter types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If virtual resource metrics are collected and managed manually in NFVI, then resource utilization can be monitored, but the management efficiency and accuracy deteriorate due to lack of automated conversion and transmission mechanisms
Solution Approach 1:
The patent introduces an intermediary conversion mechanism that automatically transforms raw virtual resource data into standardized usage metrics. This intermediary layer includes automated data collection agents, conversion rules engines, and transmission interfaces that bridge the gap between raw data and actionable metrics, thereby improving both management efficiency and measurement precision simultaneously
Solution Approach 2:
The patent replaces manual resource monitoring and metric generation with automated electronic systems. Software agents continuously collect data, algorithms automatically convert raw data to metrics according to predefined rules, and automated transmission systems deliver metrics to management platforms, eliminating manual intervention and significantly improving efficiency while maintaining high measurement accuracy
2Loss of information
If comprehensive virtual resource data is collected from all VM instances, then complete resource utilization visibility is achieved, but system complexity and data processing overhead increase
Solution Approach 1:
The patent segments the virtual resource management system into modular components: data collection agents deployed at individual compute nodes, local conversion modules that process data independently, and centralized transmission interfaces. This segmentation allows comprehensive data collection from all VM instances while distributing processing complexity across multiple independent modules, preventing system-wide complexity escalation
Solution Approach 2:
The patent implements local data conversion and processing at each compute node before transmission. Each node applies conversion rules locally to transform raw data into standardized metrics, reducing the complexity of centralized processing and enabling comprehensive resource visibility without overwhelming the central management system with raw data from all instances
Data Source
AI summary
Resource usage data of a virtual machine (VM) instance is collected by an agent to be delivered to the network function virtualization infrastructure (NFVI) for conversion to virtual resource usage metrics to be provided to a virtualized infrastructure manager (VIM) for transmission of a virtual resource usage measurement to a virtualized network function manager (VNFM). For example, an agent executing on a compute node receives virtual resource data from a compute node hypervisor. The agent provides the virtual resource data to the NFVI. The NFVI converts the virtual resource data to virtual resource usage metrics using a publishing pipeline. The NFVI uses the publishing pipeline to report the virtual resource usage metrics to a virtualized infrastructure manager. The VIM uses the data to provide virtual resource usage measurements to a virtualized network function manager (VNFM).