SASE PoP Selection via Client Performance Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The Secure Access Service Edge (SASE) model's performance is not guaranteed, as the selection of the closest point of presence (PoP) does not always result in the best application experience due to varying network conditions and SLA failures, leading to decreased user quality of experience.
Innovation Solution
A device obtains client attribute data and forms a performance model to select the optimal PoP for each client based on its attributes, using machine learning techniques to predict application experience metrics and dynamically reroute traffic to ensure the best possible performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the closest point of presence (PoP) is selected based on location and proximity, then network architecture is simplified, but application performance and user quality of experience deteriorate due to varying network conditions and SLA failures
Solution Approach 1:
The system performs preliminary actions by collecting client attribute data and forming performance models in advance, before actual application access occurs. This allows the system to predict and select the optimal PoP beforehand, rather than simply connecting to the geographically closest one, thus preventing performance issues before they arise.
Solution Approach 2:
The system changes the selection parameter from purely geographical proximity to a composite metric that includes client attributes (device type, OS, location) and predicted application experience metrics. This parameter transformation enables dynamic PoP selection that adapts to varying network conditions and client characteristics, resolving the contradiction between architectural simplicity and performance reliability.
2Loss of time
If PoP selection is based solely on geographical proximity, then connection establishment is fast and simple, but application experience metric deteriorates due to not accounting for client-specific factors and varying network conditions
Solution Approach 1:
The system pre-collects client attribute data and pre-forms performance models for multiple PoPs, so that when a client needs to connect, the optimal PoP can be quickly identified using the pre-computed models. This preliminary preparation eliminates the need for time-consuming trial-and-error connections, maintaining fast connection establishment while ensuring optimal performance.
Solution Approach 2:
The system uses feedback from collected application experience metrics to continuously refine performance models. By incorporating real-world performance data into the models, the system learns from past connections and improves future PoP selections, ensuring both speed and reliability are optimized based on actual network conditions and client characteristics.
3Device complexity
If a static PoP assignment is used, then system complexity is reduced, but adaptability to changing network conditions and client attributes deteriorates, leading to decreased performance
Solution Approach 1:
The system implements dynamic PoP selection by forming performance models that incorporate client attributes and predict application experience metrics. Unlike static assignment, this dynamic approach allows the system to adapt PoP selection based on varying network conditions, client characteristics, and application requirements, maintaining low complexity through automated model-based decision-making.
Solution Approach 2:
The system performs self-service by automatically collecting client attributes, forming performance models, and selecting optimal PoPs without requiring manual configuration or intervention. This self-service capability enables the system to adapt to changing conditions dynamically while keeping operational complexity low, as the automation handles the adaptability requirements independently.
Data Source
AI summary
In one embodiment, a device obtains client attribute data for clients of an online application that access the online application via a plurality of points of presence in a network. The device forms a performance model that models an application experience metric for the online application as a function of the client attribute data for each of the plurality of points of presence. The device selects, using the performance model, a particular point of presence from among the plurality of points of presence to be used by a particular client to access the online application, based on its client attribute data. The device causes the particular client to access the online application via the particular point of presence selected by the device using the performance model.