Border Leaf Agent Deployment for Datacenter Performance Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current deployments of application performance monitoring agents in data centers are resource-intensive and wasteful, as they are typically installed on all servers and virtual machines, providing limited visibility into network performance and requiring extensive resource usage.
Innovation Solution
Deploy application performance monitoring agents only on border leaf nodes in the datacenter, optimizing agent placement by leveraging datacenter connectivity to reduce the number of agents and tests run, while maintaining visibility into network performance from the standpoint of servers and virtual machines.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If application performance monitoring agents are deployed on all servers and virtual machines, then comprehensive network performance visibility is achieved, but resource consumption and operational complexity increase significantly
Solution Approach 1:
The patent consolidates monitoring functions by deploying agents only on border leaf nodes instead of every server and virtual machine. The border leaf nodes act as aggregation points that collect and consolidate network traffic data from multiple sources, providing comprehensive visibility while reducing the total number of agents and resource consumption by a factor of 100 or more.
Solution Approach 2:
The border leaf nodes serve as intermediary devices between internal servers/virtual machines and external networks. These intermediaries perform the monitoring function centrally, eliminating the need for agents on every endpoint while maintaining comprehensive network performance visibility through their strategic positioning in the network architecture.
2Loss of information
If application performance monitoring agents are deployed on all servers and virtual machines, then complete network performance data is collected, but deployment complexity and maintenance burden increase
Solution Approach 1:
The patent merges multiple monitoring agents into fewer border leaf node agents. This consolidation reduces deployment complexity from hundreds or thousands of individual agent installations to just a handful of border leaf nodes, while the aggregated data from these central points maintains completeness of network performance information.
Solution Approach 2:
The border leaf nodes perform multiple functions: they act as network traffic aggregation points, monitoring stations, and data consolidation hubs. This multi-functionality eliminates the need for separate agents on every server and virtual machine, simplifying deployment while maintaining comprehensive monitoring capabilities.
3Use of energy by moving object
If monitoring agents are deployed on border leaf nodes only, then resource consumption is reduced, but visibility into internal network performance may be limited
Solution Approach 1:
The border leaf nodes serve as strategic intermediaries that observe all traffic entering and leaving the data center. By positioning agents at these critical network boundary points, the system achieves comprehensive visibility into internal network performance indirectly, as all external communications must pass through these border points, providing a complete view of network performance without requiring internal agents.
Solution Approach 2:
The border leaf node agents perform monitoring actions at the network boundary before traffic enters or leaves the internal environment. This preliminary positioning allows them to capture and analyze all relevant network performance data at the source, providing comprehensive internal visibility while maintaining resource efficiency.
Data Source
AI summary
In one implementation, a device determines a plurality of routes from one or more sources in a localized network domain to external addresses via a plurality of border leaf nodes of the localized network domain. The device selects, based on a specific performance test from a particular source of the one or more sources to a particular endpoint with a corresponding external address, a particular border leaf node of the plurality of border leaf nodes that is along a utilized route from the particular source to the particular endpoint based on the corresponding external address and the plurality of routes. The device causes the specific performance test to be executed from a particular testing agent installed on the particular border leaf node to the particular endpoint.