Bandwidth Allocation Using FARIMA Forecasting for Network Traffic
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing bandwidth allocation methods for variable bit rate services in multimedia networks have low accuracy in forecasting non-stable traffic, leading to inefficient bandwidth utilization and potential network congestion.
Innovation Solution
A method that collects bandwidth historical data, obtains trend and fluctuation sequences, forecasts bandwidth fluctuations using autoregressive integrated moving average (ARIMA) models, and allocates bandwidth based on these forecasts to improve accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If resource reservation is used to guarantee QoS, then service reliability is improved, but bandwidth utilization deteriorates
Solution Approach 1:
The patent implements dynamic bandwidth allocation by using FARIMA model to forecast traffic patterns and adjust bandwidth allocation in real-time based on predicted traffic demands, replacing static resource reservation with adaptive allocation that responds to actual network conditions
Solution Approach 2:
The system changes the parameter of bandwidth allocation from fixed reserved values to dynamically adjusted values based on traffic forecasting results, allowing the allocation amount to vary according to predicted traffic needs while maintaining QoS guarantees
2Productivity
If FARIMA model is used for traffic forecast, then bandwidth utilization is improved, but forecast accuracy for non-stable traffic deteriorates
Solution Approach 1:
The patent segments the traffic forecasting process into multiple components: long-term trend analysis, short-term fluctuation prediction, and periodic pattern recognition, processing each component separately through different computational methods before combining results for final bandwidth allocation
Solution Approach 2:
The system introduces an intermediary processing layer that transforms raw traffic data into standardized sequences suitable for FARIMA modeling, applying preprocessing techniques to handle non-stationary characteristics before forecast generation
Data Source
AI summary
The present invention provides a bandwidth allocation method and device. The method includes: collecting a bandwidth historical data sequence; obtaining a bandwidth trend sequence value and a bandwidth fluctuation sequence value according to the collected bandwidth historical data sequence; obtaining a forecast sequence value of a bandwidth fluctuation sequence according to the bandwidth fluctuation sequence value; obtaining a bandwidth forecast sequence value according to the bandwidth trend sequence value and the forecast sequence value of the bandwidth fluctuation sequence; and allocating bandwidth according to the bandwidth forecast sequence value. Embodiments of the present invention are capable of improving the accuracy of bandwidth forecast, thereby allocating bandwidth more properly.