Virtual Responder for Accurate Network Performance Measurement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Network performance measurement tools, such as those used for service-level agreements, face inaccuracies due to high forwarding loads on real responders and potential failures, which can lead to unreliable packet latency and jitter measurements.

Innovation Solution

A virtual responder system that selects and provides contact information for real responders based on metrics like forwarding load, memory, and CPU usage, ensuring accurate and reliable network performance measurements by dynamically assigning the best available real responder for measurement tests.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If real responders are used for network performance measurement, then measurement functionality is provided, but measurement accuracy deteriorates due to high forwarding loads and potential failures

Engineering Contradiction:
Improvemeasurement reliabilityVSAvoidpacket latency measurement accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

A virtual responder is introduced as an intermediary between senders and real responders. The virtual responder receives measurement requests, selects appropriate real responders based on current performance metrics, and coordinates the measurement process. This mediator approach isolates senders from the complexity of multiple real responders while ensuring measurements are taken from optimally selected responders, thereby maintaining measurement accuracy even when individual real responders experience high loads or failures.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system continuously monitors performance metrics of real responders and uses this feedback to dynamically select the most suitable real responder for each measurement request. The virtual responder polls real responders for metrics such as forwarding load, memory availability, and CPU usage, then selects responders based on current conditions. This feedback mechanism ensures that measurements are consistently obtained from responders in optimal states, improving measurement reliability and accuracy.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If multiple real responders are monitored for optimal selection, then measurement accuracy improves, but system complexity increases

Engineering Contradiction:
Improvemeasurement accuracyVSAvoidvirtual responder system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

Multiple real responders are merged into a single virtual responder entity from the sender's perspective. The virtual responder consolidates the functionality of multiple real responders and presents a unified interface to senders. This merging approach maintains the benefits of having multiple monitored responders for accurate measurement selection while simplifying the system architecture by providing a single point of contact, thereby reducing operational complexity despite the underlying complexity of managing multiple real responders.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentUS8054750B2Virtual responder for the auto-discovery of a real responder in a network performance test
Publication Date: 2011.11.08 CISCO TECHNOLOGY INC
  • US8054750B2 patent drawing
  • US8054750B2 patent drawing
  • US8054750B2 patent drawing

AI summary

Particular embodiments provide a virtual responder that can be contacted by any number of senders. When a sender wants to perform a measurement test with a real responder, the sender sends a control message to a virtual responder. The virtual responder is configured to determine a real responder based on real responder selection metrics. The virtual responder may poll the real responders at certain times to receive the real responder selection metrics. The virtual responder then selects a real responder from a plurality of real responders based on the real responder's selection metrics. The virtual responder then sends contact information for the selected real responder to the sender device. This allows the sender device to contact the selected real responder to perform the measurement test with a real responder.