Predictive SASE PoP Selection for SLA Compliance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The Secure Access Service Edge (SASE) model faces challenges in ensuring consistent performance and meeting service level agreements (SLAs) due to the variability in performance of points of presence (PoPs) over time, which can lead to decreased user experience and SLA violations.
Innovation Solution
A predictive system that collects telemetry data from edge routers sending probes to multiple PoPs, uses machine learning to forecast potential SLA violations, and dynamically selects the optimal PoP for accessing cloud-hosted applications to prevent such violations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the closest PoP is selected based on location and proximity to the edge device, then the network architecture is simplified, but the performance and quality of experience may decrease over time due to PoP performance variability
Solution Approach 1:
The system dynamically selects PoPs based on real-time performance predictions rather than using a static closest-PoP assignment. The predictive model continuously evaluates multiple PoPs and adjusts selections based on forecasted performance, allowing the system to adapt to changing conditions while maintaining simplified architecture.
Solution Approach 2:
The system performs preliminary performance evaluation and prediction before actual traffic routing. By forecasting PoP performance in advance using machine learning models on historical telemetry data, the system proactively selects optimal PoPs before performance degradation occurs, preventing SLA violations rather than reacting to them.
2Ease of operation
If a single closest PoP is used for each edge device, then the configuration is simplified, but performance variability and SLA violations increase over time
Solution Approach 1:
The system maintains a universal pool of available PoPs that can serve multiple edge devices. Instead of dedicating a single PoP to each device, the same set of PoPs can be dynamically selected for different devices based on real-time performance predictions, providing both operational simplicity and performance consistency.
Solution Approach 2:
The system implements feedback loops where telemetry data from edge devices and PoPs is continuously collected, analyzed by machine learning models, and used to adjust PoP selections. This closed-loop control ensures that performance issues are detected and corrected through dynamic reselection, maintaining consistent application performance.
3Extent of automation
If PoP performance is not monitored and predicted, then the system operation is simpler, but SLA violations and user experience degradation occur
Solution Approach 1:
The system performs self-monitoring and self-optimization through automated machine learning models that continuously analyze telemetry data and make PoP selection decisions without manual intervention. The predictive analytics engine autonomously identifies performance trends and adjusts routing decisions, eliminating the need for manual PoP management while ensuring SLA compliance.
Solution Approach 2:
The system replaces manual PoP selection and monitoring mechanisms with automated machine learning-based predictive analytics. Instead of human operators manually configuring and monitoring PoP performance, intelligent algorithms automatically analyze telemetry data, predict performance issues, and make optimal PoP selections, substituting mechanical/manual processes with intelligent automation.
Data Source
AI summary
In one embodiment, a device obtains telemetry data that results from an edge router sending probes to a cloud-hosted application via a plurality of points of presence. The device makes, based on the telemetry data, predictions as to whether use of each of the plurality of points of presence by the edge router to access the cloud-hosted application will result in a violation of a service level agreement. The device selects, based on the predictions, a particular point of presence from among the plurality of points of presence that the edge router should use to access the cloud-hosted application during a time window. The device causes the edge router to access the cloud-hosted application via the particular point of presence during the time window.