Time-Based Traffic Routing for Datacenter Load Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional load aware traffic management approaches in datacenters are reactive and resource-intensive, failing to effectively prevent server overload and performance issues due to the need for continuous health monitoring and data analysis.
Innovation Solution
Implementing a time-based traffic management system that utilizes scheduled real-world events to proactively adjust traffic routing configurations, eliminating the need for health monitoring and reducing resource expenditure by predicting and managing traffic loads based on predictable patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If load aware traffic management is implemented to control endpoint overload, then server reliability is improved, but system complexity and resource consumption increase due to continuous health monitoring and data analysis
Solution Approach 1:
The patent applies preliminary action by proactively routing traffic away from endpoints during identified high-load time periods before overload occurs. Instead of reactively monitoring and responding to load conditions, the system pre-determines time periods with historically high traffic patterns and automatically adjusts routing configurations in advance, eliminating the need for continuous health monitoring while maintaining server reliability.
Solution Approach 2:
The patent extracts the continuous health monitoring and real-time data analysis components from the traffic management system. By separating these resource-intensive functions and replacing them with time-based routing rules, the system maintains load control capabilities while significantly reducing computational overhead and system complexity.
2Measurement precision
If continuous health monitoring is performed to identify overload situations, then traffic routing accuracy is improved, but computing resources and time are excessively consumed
Solution Approach 1:
The patent implements periodic action by analyzing traffic patterns at scheduled intervals to identify time periods with high traffic patterns, rather than performing continuous real-time monitoring. The system periodically reviews historical traffic data, determines high-load time periods, and applies routing configurations accordingly, significantly reducing computing resource consumption while maintaining adequate routing accuracy.
Solution Approach 2:
The patent replaces expensive continuous monitoring operations with simpler, less resource-intensive time-based routing rules. Instead of continuously collecting and analyzing detailed health metrics from endpoints, the system uses pre-determined time period configurations that require minimal computational resources to evaluate and enforce.
3Productivity
If reactive load management is used to redirect traffic after overload begins, then endpoint capacity utilization is maximized, but performance degradation occurs before traffic can be redirected
Solution Approach 1:
The patent applies preliminary action by proactively identifying time periods with historically high traffic patterns and pre-configuring routing rules before overload occurs. This allows traffic to be redirected in advance to alternative endpoints, preventing performance degradation while maintaining high overall capacity utilization across the endpoint fleet.
Solution Approach 2:
The patent implements beforehand cushioning by creating a buffer against potential overload conditions through pre-determined routing configurations. When high-load time periods are identified, the system prepares alternative routing paths in advance, cushioning the system against performance degradation before it can occur during peak traffic periods.
Data Source
AI summary
The techniques described herein enable the use of a time factor for traffic management and routing. A system is configured to analyze traffic for a service over a period of time and identify (e.g., learn) traffic patterns that reflect a substantial effect on traffic during a particular real-world event. Using the traffic patterns identified via the analysis, the system can provide valuable time-based traffic information to service providers. A service provider can then create a predefined time-based profile that is used by a traffic manager to switch from a current traffic routing configuration to a different traffic routing configuration that better accommodates an expected traffic load for various endpoints. The predefined time-based profile specifies a scheduled time at which the switch is to occur, and this scheduled time can correspond to a start time for a real-world event that is known to cause an increase or decrease in traffic.