Predictive autoscaler for cellular communication networks
A predictive autoscaler using machine learning models in 5G networks addresses inefficient scaling by anticipating future loading, optimizing resource allocation and improving network performance through proactive scaling of microservices.
WO2026117239A1PCT designated stage Publication Date: 2026-06-04RAKUTEN SYMPHONY INC +1
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- RAKUTEN SYMPHONY INC
- Filing Date
- 2024-11-27
- Publication Date
- 2026-06-04
AI Technical Summary
Technical Problem
Existing cellular communication networks struggle with inefficient scaling of software components due to reactive scaling methods that do not anticipate future loading demands, leading to potential overloading or underutilization.
Method used
Implementing a predictive autoscaler system that uses trained machine learning models to process network signals, extract relevant features, and predict future loading of microservices, enabling proactive scaling of software components in 5G networks.
Benefits of technology
The system enables efficient and timely scaling of microservices based on predicted loading, optimizing resource utilization and reducing the risk of overloading or underutilization, thereby enhancing network performance and efficiency.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure US2024057745_04062026_PF_FP_ABST
Abstract
A cellular communication network includes one or more computing devices and configured to process connections to user equipment, each computing device of the one or more computing devices executing one or more software components. Each computing device of the one or more computing devices is configured to: receive a plurality of signals from the one or more microservices; process the plurality of signals with a machine learning model to obtain a predicted loading by the user equipment; and scale up or scale down at least one of the microservices according to the predicted loading.
Need to check novelty before this filing date? Find Prior Art