Server Load Prediction Using Periodic Trends for Peak Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current server load prediction models in cloud computing are inaccurate, particularly at peak loads, leading to potential service defaults and increased costs due to the need for redundant resources to ensure reliability.
Innovation Solution
A deep learning-based method that incorporates periodic information and long-term trend analysis using an encoder-decoder architecture with LSTM and multi-headed self-attention mechanisms to improve peak load prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current prediction models are used for server load with drastic changes, then general prediction capability is maintained, but prediction accuracy at peak load deteriorates
Solution Approach 1:
The prediction model is segmented into two specialized components: a model specifically optimized for peak load prediction and another for normal load prediction. This segmentation allows each model to be tailored to its specific prediction scenario, with the peak load model focusing exclusively on capturing peak characteristics without being constrained by normal load patterns, thereby resolving the contradiction between general prediction capability and peak prediction accuracy.
Solution Approach 2:
The system performs preliminary identification of peak load periods before making predictions. By detecting and marking peak periods in advance using historical data analysis and pattern recognition, the system can then route these identified peak periods to the specialized peak prediction model, ensuring that peak loads receive dedicated prediction attention and improving both accuracy and reliability during critical periods.
2Reliability
If redundant computing resources are preset to ensure high reliability, then service reliability is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The system implements a feedback mechanism where accurate peak load predictions inform dynamic resource allocation decisions. By continuously monitoring prediction accuracy and comparing predicted versus actual peak loads, the system learns and adjusts its predictions over time, enabling more precise resource provisioning that maintains reliability while reducing unnecessary redundancy and improving utilization efficiency.
Solution Approach 2:
The system dynamically changes resource allocation parameters based on prediction confidence and historical accuracy. During periods of high prediction confidence and proven accuracy, the system can reduce redundant resource provisioning while maintaining reliability standards. This parameter adjustment allows the system to optimize the balance between reliability and resource utilization efficiency by adapting resource levels to actual predicted needs rather than maintaining static redundancy.
Data Source
AI summary
The present invention discloses a server load prediction method based on deep learning, collecting the trend change of server load long series, and utilizing server load periodic information to establish a deep neural network prediction model to optimize peak load prediction. The present invention provides a method for improving the accuracy of neural network prediction by combining periodic information, long-term trend information, and short-term time series information, and demonstrates superiority over traditional methods in the peak load section. The method of the present invention can effectively improve prediction accuracy, provide more accurate scheduling and evacuation decision-making basis for cloud service providers, thereby reducing the redundant equipment required to ensure high reliability, reducing the operating costs of cloud service providers, and reducing the rental expenses of cloud service tenants.