Virtualized Application Telemetry via Kernel-Level Resource Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for reporting operating health in virtualized environments using network connections incur latency, leading to delayed resource adjustments by orchestrators, and lack reliable methods for monitoring CPU frequency and cache usage in multi-VNF environments.
Innovation Solution
Implementing a service assurance agent with kernel-level privileges to directly access and adjust CPU/core resources based on real-time telemetry data from virtualized environments, using a hardware virtualization framework to expose performance indicators directly to the orchestrator.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If network connection sockets are used to report operating health telemetry from virtual environments to orchestrators, then telemetry reporting can be implemented, but latency increases and resource adjustment responsiveness deteriorates
Solution Approach 1:
The patent extracts the telemetry reporting mechanism from the network layer and places it directly in the kernel space, eliminating network traversal latency. The service assurance agent reads telemetry data directly from kernel data structures and exposes it to orchestrators through local interfaces, bypassing the network socket pathway that causes delays.
Solution Approach 2:
The service assurance agent acts as an intermediary component with kernel-level privileges that mediates between the virtual environment telemetry sources and the orchestrator. It directly accesses kernel data structures and provides telemetry data to orchestrators without requiring network communication, thus reducing latency while maintaining reliable reporting.
2Reliability
If custom-made software is used to report operating health in virtualized environments, then telemetry can be collected, but the system complexity increases and integration difficulty arises
Solution Approach 1:
The service assurance agent is designed as a universal kernel-level component that can collect telemetry data from multiple virtual environments and applications simultaneously. It provides a standardized interface for orchestrators to access telemetry data, eliminating the need for separate custom software solutions for each virtual environment and reducing overall system complexity.
3Productivity
If CPU resources are allocated to handle network processing operations, then network processing capability is improved, but the ability to monitor and adjust resources based on real-time performance deteriorates due to latency
Solution Approach 1:
The patent implements a feedback mechanism where the service assurance agent continuously monitors CPU resource utilization and performance metrics in real-time and exposes this information to orchestrators without network latency. Orchestrators can immediately adjust resource allocation based on current performance conditions, creating a closed-loop feedback system that optimizes network processing capability while maintaining responsive resource management.
Data Source
AI summary
Examples describe techniques to expose application telemetry in a virtualized execution environment. Examples include a plurality of application executing within the virtualized execution environment writing telemetry data to a memory associated with virtual devices of a hardware device. Examples also include an orchestrator to read the telemetry data from the memory and use the telemetry data to make resource allocation decisions.