Network Path Selection Using Historical Performance Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network path selection methods often rely on recent snapshots of performance metrics, failing to consider historical performance and SLA compliance, leading to suboptimal routing of data flows and increased retransmissions, which wastes bandwidth and reduces application performance.
Innovation Solution
Implement a method that monitors and stores historical network performance metrics for multiple paths, using machine learning algorithms to determine aggregate historical performance, and route data flows based on the path most likely to satisfy Service Level Agreements (SLAs) by considering factors like jitter, latency, loss, cost, and carrier reputation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If network path selection relies on recent snapshots of performance metrics, then the routing decision is simple and fast, but the reliability and SLA compliance of the selected path deteriorates
Solution Approach 1:
The system performs preliminary actions by continuously monitoring and storing historical performance metrics for multiple network paths before actual data flow transmission. This pre-collection of performance data enables more reliable path selection decisions without adding complexity to the actual routing decision process.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring network performance metrics and using this information to dynamically adjust path selection. Historical performance data is fed back into the path selection algorithm, creating a closed-loop system that improves reliability through learned patterns while maintaining manageable complexity through automated decision-making.
2Reliability
If historical performance data is collected and analyzed for path selection, then the SLA compliance improves, but the bandwidth waste due to retransmissions worsens
Solution Approach 1:
The system performs preliminary analysis of historical performance data to identify optimal paths before data flow transmission begins. By pre-processing performance metrics and establishing performance profiles for different paths, the system makes accurate routing decisions that minimize retransmissions and bandwidth waste from the outset.
Solution Approach 2:
The network system serves itself by automatically monitoring, analyzing, and selecting paths based on historical performance data without external intervention. This self-service capability ensures consistent SLA compliance while minimizing bandwidth waste through automated, data-driven path selection that adapts to changing network conditions.
3Measurement precision
If multiple performance metrics are monitored for path selection, then the decision accuracy improves, but the processing time and complexity increases
Solution Approach 1:
The system performs preliminary monitoring and aggregation of multiple performance metrics continuously, storing historical data for latency, jitter, packet loss, and other parameters. This pre-aggregation of multi-dimensional performance data enables fast, accurate path selection decisions without requiring real-time analysis of all metrics simultaneously.
Solution Approach 2:
The system transforms multiple raw performance metrics into aggregated performance scores or rankings that capture the essential characteristics of each path. By changing parameters from individual metric values to composite performance indicators, the system maintains measurement precision while reducing the computational burden of path selection.
Data Source
AI summary
A method may include monitoring a network performance metric for multiple paths to a destination through a network, and storing historical performance data for the paths. The method may also include receiving a data flow directed to the destination, where the data flow may be subject to a network performance agreement. The method may additionally include determining aggregate historical performances for the paths, and comparing the aggregate historical performances for the paths. The method may also include, based on the comparison of the aggregate historical performances, routing the data flow through the network.