Network Controller Performance Monitoring Framework
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current frameworks for data centers lack effective tools to optimize network performance monitoring in virtualized and non-virtualized environments, particularly in telecom-oriented services with high compute bias and lax latency requirements.
Innovation Solution
A method and apparatus for network performance monitoring that involves a network controller sending requests for performance data to a performance database manager, utilizing a scalable framework to collect and manage performance metrics across multiple service monitoring servers, enabling intelligent load balancing and traffic steering based on performance data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a centralized framework is used to optimize cluster resources (CPU, Memory, Storage, Power), then resource optimization is improved, but network performance monitoring capability deteriorates
Solution Approach 1:
The framework is segmented into distinct functional modules: resource management components (DRS, Vscheduler, MEMB) for CPU/memory/storage/power optimization, and network monitoring components (performance data collectors, network performance metrics) for connectivity monitoring. This segmentation allows each module to specialize in its domain while working within the unified framework architecture.
Solution Approach 2:
The performance monitoring framework is designed to be universal across both virtualized and non-virtualized environments. It can monitor multiple resource types (CPU, memory, storage, power, network) through a common architecture that adapts to different deployment scenarios, making it applicable to diverse data center configurations.
2Measurement precision
If application performance metrics are collected requiring Guest OS cooperation, then measurement precision is improved, but system complexity deteriorates
Solution Approach 1:
Performance collectors are introduced as intermediary components that run within the Guest OS and act as mediators between the application layer and the host monitoring infrastructure. These collectors gather precise application performance metrics (socket operations, buffer usage, I/O operations) and export them through standardized interfaces, reducing the complexity burden on both the Guest OS and the host monitoring system.
Solution Approach 2:
The framework implements feedback mechanisms where performance data is continuously collected from applications, processed by the performance database manager, and used to generate actionable insights. This feedback loop enables dynamic adjustment of resource allocation and performance optimization strategies based on real-time application metrics.
Data Source
AI summary
A method and apparatus for providing network performance monitoring is disclosed. At least one application is run on a network controller. The at least one application running on the network controller sends a request for performance data information to a performance database manager (PDM). A response to the request for performance data information is received from the PDM.A method for determining a service path using network performance monitoring data is disclosed. Performance data is determined for one or more applications. A service path is assigned by steering traffic to a particular instance of each of the one or more applications using the performance data. The steered traffic is forwarded to a next application based on a current position along a service path.