Adaptive Traffic Engineering via Predicted Demand
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current SDN-WAN technologies face inefficiencies due to static traffic engineering approaches that do not adapt to changing traffic patterns, leading to over-provisioning and lack of real-time feedback in packet optical communication networks.
Innovation Solution
Implementing a system that monitors packet traffic, learns demand between devices, predicts traffic demand based on workload history, and dynamically adjusts data paths using OpenFlow protocols to establish optimal network paths, enabling real-time monitoring and adaptation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If static traffic engineering paths are provisioned at initialization, then network management is simplified, but the system cannot adapt to changing traffic patterns leading to operational inefficiencies
Solution Approach 1:
The patent implements dynamic traffic engineering by continuously monitoring traffic volumes on network connections and automatically re-provisioning paths based on observed traffic patterns. The system transitions from static initialization-based path provisioning to dynamic adaptation, where paths are adjusted in real-time based on monitored traffic demands, resolving the contradiction between ease of operation and adaptability.
Solution Approach 2:
The system establishes a feedback loop by monitoring traffic volumes on network connections and using this information to dynamically re-provision paths. The controller observes actual traffic patterns and adjusts traffic engineering decisions accordingly, enabling the system to adapt to changing conditions while maintaining automated management through the feedback mechanism.
2Reliability
If paths are provisioned to account for worst-case scenario, then reliability is improved, but network resource utilization becomes inefficient due to over-provisioning
Solution Approach 1:
The system dynamically adjusts path provisioning based on actual monitored traffic volumes rather than static worst-case assumptions. By continuously observing traffic patterns and adapting paths in real-time, the system maintains reliable service delivery while optimizing resource utilization to match actual demand, eliminating the need for persistent over-provisioning.
Solution Approach 2:
The system changes traffic engineering parameters dynamically based on monitored traffic conditions. Instead of fixed provisioning parameters set at initialization, the controller adjusts path selections and traffic distribution parameters in response to observed traffic volumes, allowing the system to maintain reliability while adapting resource allocation to actual usage patterns.
3Device complexity
If static traffic engineering is used, then device complexity is reduced, but there is no feedback mechanism to account for changing traffic patterns
Solution Approach 1:
The patent implements a feedback mechanism where the controller monitors traffic volumes on network connections and uses this information to dynamically adjust path provisioning. This feedback loop captures traffic pattern information that would otherwise be lost in static systems, enabling adaptive decision-making while maintaining relatively simple device implementations through centralized control.
Data Source
AI summary
Systems and methods for adaptive and automated traffic engineering of data transport services may include learning the demand between devices and data paths based on application workloads, prediction of traffic demand and paths based on the workload history, provisioning and management of data paths (i.e. network links) based on the predicted demand, and real-time monitoring and data flow adaptation. Systems and methods for adaptive and automated traffic engineering of data transport services may also include learning the variation of traffic (data flow in the network) on various links (data paths) of the network topology using historical data (e.g. a minute, an hour, a day, or a week of data), predicting the data flow pattern for a time interval, and provisioning the services to steer data to meet the application requirements and other network wide goals (e.g., load balancing).