Network Adapter Packet Coloring for Application Telemetry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network analysis using telemetry data struggles to distinguish between different application programs and execution states, leading to aggregated data that is difficult to analyze for performance degradation causes.
Innovation Solution
Network devices generate telemetry data based on application-level metadata marked in packets, allowing for selective tagging and analysis of data related to specific applications and execution states.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network devices collect and report telemetry data for all packets, then comprehensive network performance data is obtained, but the data becomes aggregated and difficult to analyze for specific application performance degradation
Solution Approach 1:
The patent segments telemetry data collection by applying packet coloring markers that divide network traffic into distinct application-specific groups. Each marker identifies packets belonging to particular application programs or execution states, allowing the telemetry system to segment aggregated data into application-specific subsets for precise analysis without losing application-level information.
2Measurement precision
If packet markers are added to identify application-level metadata, then precise application identification is achieved, but packet processing complexity increases
Solution Approach 1:
The patent applies preliminary action by marking packets with application-level metadata at the source network device before packets traverse the network. This pre-marking approach eliminates the need for complex analysis at intermediate devices, as the identification work is completed in advance during packet generation or initial processing.
Solution Approach 2:
The patent extracts application-level metadata from complex application execution states and encapsulates it into compact packet markers. This extraction process separates the essential identification information from the full application context, enabling efficient processing while maintaining precise application identification capability.
Data Source
AI summary
An apparatus includes an interface and a processor. The interface is to receive telemetry data items obtained from one or more network adapters and from one or more network switches of a network, the telemetry data items pertaining to respective packets and including respective identifiers of flows or applications with which the packets are associated. The processor is to analyze the telemetry data items using the identifiers.