Bandwidth Scheduling for Data Center Traffic Peaks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In public cloud data centers, physical lines in the data center interconnection network experience low utilization due to significant differences between traffic peaks and valleys, leading to wasted resources and potential packet loss during peak times.
Innovation Solution
A bandwidth scheduling method that allocates bandwidth based on historical data to efficiently manage non-real-time traffic, balancing resource usage by predicting and scheduling bandwidth requirements across different time periods, ensuring stable occupation of total bandwidth and reducing the difference between peak and valley traffic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If physical line bandwidth is expanded based on peak traffic, then packet loss during transmission is avoided, but physical line resources are wasted due to low utilization during traffic valleys
Solution Approach 1:
The system performs preliminary bandwidth allocation by predicting future bandwidth requirements based on historical data and traffic patterns. The bandwidth scheduling module allocates bandwidth in advance for non-real-time traffic during traffic valleys before peak periods occur, ensuring that physical lines are fully utilized without requiring expansion for peak traffic alone.
Solution Approach 2:
The system dynamically adjusts bandwidth allocation based on real-time traffic conditions and predicted future needs. The bandwidth scheduling module continuously monitors traffic patterns and modifies allocation strategies, transitioning from static peak-based expansion to dynamic allocation that adapts to varying traffic demands throughout different time periods.
2Reliability
If physical line bandwidth is expanded to handle peak traffic, then transmission reliability is improved, but device complexity increases due to managing expanded capacity
Solution Approach 1:
The system implements feedback mechanisms where the bandwidth scheduling module continuously receives traffic status information, analyzes patterns, and adjusts bandwidth allocation accordingly. This closed-loop control enables the system to maintain transmission reliability through intelligent scheduling rather than physical expansion, reducing the complexity of managing expanded infrastructure.
Solution Approach 2:
The system changes operational parameters by shifting from fixed bandwidth allocation based on peak traffic to variable allocation based on predicted traffic patterns. The bandwidth scheduling module adjusts allocation parameters dynamically, allowing the same physical infrastructure to handle varying loads efficiently without requiring additional physical lines or complex management of expanded capacity.
3Productivity
If bandwidth is allocated based on historical information for non-real-time traffic, then physical line utilization is improved, but bandwidth allocation complexity increases
Solution Approach 1:
The bandwidth scheduling module operates autonomously by automatically analyzing historical bandwidth information, predicting future requirements, and allocating bandwidth without manual intervention. The system serves itself by continuously optimizing allocation based on learned patterns, improving physical line utilization while managing allocation complexity through automation rather than manual processes.
Solution Approach 2:
The system replaces manual or mechanical bandwidth allocation methods with automated computational analysis. The bandwidth scheduling module uses algorithms to process historical data and generate allocation decisions, substituting complex manual planning with automated systems that handle the complexity internally while presenting simple interfaces for resource management.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
Embodiments of this application disclose a bandwidth scheduling method, a traffic transmission method, and a related product. The bandwidth scheduling method includes: receiving a bandwidth request sent by a data center, where the bandwidth request includes bandwidth required to transmit non-real-time traffic; allocating, based on historical bandwidth information, the bandwidth required by the data center to transmit the non-real-time traffic in a future time period, where the historical bandwidth information is used to predict occupation of total bandwidth by the data center in a region in which the data center is located at each moment in the future time period; and sending a bandwidth response to the data center, where the bandwidth response includes an allocation result. Embodiments of this application help improve utilization of physical line resources in each region.