Automated Traffic Routing with Dynamic Threshold Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional systems for automatic traffic routing rely on single monitoring systems and limited error thresholds, leading to inefficiencies and downtime when issues arise, and do not effectively monitor client impact.
Innovation Solution
A system utilizing multiple monitoring systems and a machine learning model (MLM) to iteratively monitor application traffic, transmit queries, and determine if thresholds are exceeded, allowing for dynamic adjustment of traffic flows between new and previous application stacks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single monitoring system is used for traffic routing, then the system complexity is reduced, but the reliability and detection capability deteriorate when issues arise
Solution Approach 1:
The patent divides the monitoring function into multiple independent monitoring systems (first monitoring system, second monitoring system, etc.), each monitoring different aspects of application traffic. This segmentation allows the system to detect various types of issues through different monitoring perspectives, improving overall reliability without requiring a single overly complex monitoring system.
Solution Approach 2:
Each monitoring system is assigned specific monitoring tasks and thresholds tailored to particular aspects of traffic flow. For example, different monitoring systems monitor different error thresholds or traffic characteristics, allowing each component to be optimized for its specific function rather than requiring a monolithic complex system.
2Device complexity
If fixed error thresholds are used for traffic diversion, then the control logic is simplified, but the adaptability to changing traffic conditions deteriorates
Solution Approach 1:
The patent implements dynamic threshold adjustment where monitoring systems continuously evaluate traffic conditions and automatically adjust error thresholds and routing decisions. The system transitions from static fixed thresholds to dynamic adaptive thresholds that respond to changing traffic patterns, application performance, and error rates, improving adaptability while maintaining manageable control logic through automated feedback loops.
Solution Approach 2:
The monitoring systems continuously collect traffic data and feed this information back to the routing control mechanism. This feedback loop enables the system to automatically adjust routing decisions based on real-time performance metrics, allowing adaptability to changing conditions without requiring complex manual control logic.
3Object-affected harmful factors
If traffic is diverted back to previous stack upon issue detection, then client impact is reduced, but productivity and deployment efficiency deteriorate due to repeated rollbacks
Solution Approach 1:
The patent implements partial traffic routing where only a portion of traffic is diverted to the previous stack when issues are detected, while another portion continues to the new stack. This partial action allows the system to protect affected clients by routing their traffic to the stable previous version while still allowing other traffic to continue testing the new deployment, thus reducing client impact without completely halting productivity gains from the new stack.
Solution Approach 2:
The monitoring and routing system automatically detects issues and adjusts traffic allocation without requiring manual intervention. The system self-corrects by diverting traffic when problems are detected and can self-restore by increasing traffic to the new stack when issues are resolved, eliminating the need for manual rollback operations and improving deployment efficiency while protecting clients.
4Reliability
If multiple monitoring systems with dynamic thresholds are implemented, then reliability and adaptability are improved, but device complexity increases
Solution Approach 1:
The patent designs monitoring systems that perform multiple functions: they monitor traffic flow, detect errors, evaluate performance metrics, and control routing decisions. By making each monitoring system multi-functional, the patent reduces the need for separate specialized systems for each function, thereby improving reliability and adaptability while controlling overall system complexity through universal components.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed embodiments may include a system for automatic traffic routing. The system may receive data associated with an application, and may assign a first portion of application traffic to be routed to a new stack, and a second portion of the application traffic to be routed to a previous stack. The system may continuously monitor the application traffic by utilizing a plurality of monitoring systems. The system may utilize a machine learning model (MLM) to iteratively transmit one or more respective monitoring queries to each of the plurality of monitoring systems, and determine whether an aspect of a monitoring query exceeds a threshold. Responsive to such determination, the system may assign a third portion of the application traffic to be routed to the new stack, and a fourth portion of the application traffic to be routed to the previous stack.