L3 L4 Network Elements Inspecting L7 Tracing Information

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Distributed tracing systems and network infrastructure are mutually agnostic, missing opportunities for network optimization and visibility, as network infrastructure cannot inspect or utilize distributed tracing information, and vice versa.

Innovation Solution

Extending the capabilities of network elements to inspect, extract, and complement tracing information added to L7 flows by mapping L7 tracing information to L3 or L4 packets, allowing L3/L4 network elements to make optimized network decisions based on this information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If distributed tracing systems use HTTP headers to propagate trace context, then tracing information can be collected across microservices, but network infrastructure cannot inspect or utilize this tracing information

Engineering Contradiction:
Improvetracing information visibilityVSAvoidnetwork infrastructure capability
Core Design Contradiction:
Loss of informationVSAdaptability or versatility

Solution Approach 1:

The patent introduces an intermediary mechanism that translates L7 tracing information from HTTP headers into L3/L4 packet markings. This intermediary translation layer enables network infrastructure to access tracing information without directly inspecting L7 headers, resolving the contradiction between maintaining L7 abstraction and enabling network visibility.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent moves tracing information from the L7 application layer dimension to the L3/L4 network layer dimension by embedding trace context in packet markings. This dimensional transition allows network elements to access tracing information at the network layer without breaking L7 protocol independence.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Productivity

If network infrastructure remains agnostic to distributed application tracing, then network elements can operate independently, but opportunities for network optimization based on tracing information are missed

Engineering Contradiction:
Improvenetwork optimization capabilityVSAvoidnetwork element complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-marking packets with tracing information at the L3/L4 layer before they reach network elements. This allows network elements to make optimized decisions based on tracing information without adding complex L7 inspection capabilities, as the tracing data is already available at the network layer.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If distributed tracing systems collect network-related information from application perspective, then application-level metrics can be obtained, but actual first-hand network information is missing

Engineering Contradiction:
Improvenetwork information accuracyVSAvoidfirst-hand network information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent implements feedback by having L3/L4 network elements observe actual network conditions and correlate them with L7 tracing information. This creates a feedback loop where network elements can provide accurate, first-hand network measurements back to the distributed tracing system, improving measurement precision with actual network data rather than just application-level observations.

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS12010001B2Extending distributed application tracing for network optimizations
Publication Date: 2024.06.11 CISCO TECHNOLOGY INC
  • US12010001B2 patent drawing
  • US12010001B2 patent drawing
  • US12010001B2 patent drawing

AI summary

Techniques for extending network elements to inspect, extract, and complement tracing information added to L7 flows by application distributed tracing systems. The techniques may include receiving a Layer-7 (L7) message of an L7 flow associated with a distributed application and determining that the L7 message includes tracing information. In some examples, the tracing information may be mapped to a marking that is to be included in a Layer 3 (L3) or Layer-4 (L4) packet carrying the L7 message, and the L3 or L4 packet including the marking may be sent to an L3 or L4 network element. In some examples, the L3 or L4 network element may be configured to utilize the marking to determine a network decision for the L3 or L4 packet.