Network Service Level Prediction via Virtual Link Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network capacity planning systems face inaccuracies and inconsistencies when modeling network communication constraints for remote locations, relying heavily on the integrity of the model used, which can lead to unreliable evaluations of service levels.
Innovation Solution
A method and system that store a service level objective for a network condition on a first link, matching it on a second link connecting to a remote location, allowing for the prediction of whether the service level of transactions meets the objective by varying multiple network conditions while keeping others constant, including bandwidth, latency, packet loss, and quality of service mechanisms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If network capacity planning systems use analytical and/or discreet event simulation tools to model network communication constraints, then they can evaluate service levels at remote locations, but the measurements are subject to inaccuracies and inconsistencies due to model integrity issues
Solution Approach 1:
The patent creates a virtual copy of the remote network location and its network conditions in the laboratory environment. By replicating the actual network conditions (bandwidth, latency, packet loss, etc.) on a local testbed, the system eliminates the need to rely on potentially inaccurate analytical models while maintaining measurement accuracy. The virtual copy allows direct observation and measurement of real network behavior under controlled conditions.
Solution Approach 2:
The patent performs preliminary measurements and characterizations of network conditions at the remote location before conducting capacity planning evaluations. By pre-capturing actual network performance data and using it to configure the laboratory testbed, the system ensures that subsequent measurements are based on real-world conditions rather than theoretical models, thereby improving both accuracy and reliability.
2Adaptability or versatility
If the system varies multiple network conditions concurrently to test service levels, then it can evaluate complex network scenarios, but it becomes difficult to determine which specific condition caused service level failures
Solution Approach 1:
The patent segments the network conditions into individual test cases. Instead of varying multiple conditions simultaneously, the system conducts separate experiments where only one network condition is varied at a time while others remain constant. This segmentation allows clear identification of which specific condition causes service level failures, while still enabling comprehensive evaluation of complex network scenarios through multiple individual tests.
Solution Approach 2:
The patent applies different network conditions to different aspects of the test configuration. Each network condition (bandwidth, latency, packet loss, etc.) is applied independently to specific components of the network model, allowing precise control and observation of their individual effects on service levels while maintaining the ability to test complex overall scenarios.
Data Source
AI summary
A system and method to store a service level objective, produce one or more network conditions on a first network link where such network condition matches the state of such network condition on a second network link, and where the second network link connects to a remote network location, execute a transaction of an application over the first network link and predict if a service level of the transaction of the application over the second network link reaches the stored service level objective.