SDN Controller Learning Capability for Dynamic Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current radio access networks face inefficiencies in resource utilization due to varying service requirements and growing mobile traffic, lacking intelligent data collection and dynamic resource allocation mechanisms to optimize bandwidth and latency for diverse services.
Innovation Solution
An enhanced software-defined network (SDN) controller with a learning capability is implemented, collecting data on a per-flow basis from the radio access network to apply learning algorithms for determining network usage requirements and allocating resources dynamically based on bandwidth and latency needs, thereby enhancing network resource management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional radio access network resource allocation is used, then device complexity is low, but network resource utilization efficiency deteriorates due to inability to adapt to varying service requirements
Solution Approach 1:
The patent segments the SDN controller into multiple functional modules: a data collection module that gathers network usage data, a learning algorithm module that processes the data, and a resource allocation module that executes decisions. This modular segmentation enables complex learning functions while maintaining manageable system architecture and improving network resource utilization through intelligent adaptation to varying service requirements
Solution Approach 2:
The patent introduces a learning algorithm as an intermediary between raw network usage data and resource allocation decisions. This intermediary processes collected data to identify patterns and generate optimized allocation strategies, thereby improving network resource utilization efficiency without requiring direct complex interactions between data collection and allocation functions
2Adaptability or versatility
If static resource allocation is used, then system complexity is low, but adaptability to changing service demands deteriorates
Solution Approach 1:
The patent implements dynamic resource allocation by continuously collecting network usage data and applying learning algorithms that adapt to changing service demands. The system dynamically adjusts resource allocation based on learned patterns, enabling high adaptability to varying traffic conditions while managing complexity through automated learning processes rather than manual configuration
Solution Approach 2:
The patent establishes a feedback loop where network usage data is continuously collected, processed by learning algorithms, and used to adjust resource allocation decisions. This feedback mechanism enables the system to adapt to changing service demands automatically, improving versatility while the automated nature of the feedback process manages the complexity of the adaptation mechanism
3Measurement precision
If per-flow data collection is implemented, then measurement precision of network usage requirements improves, but loss of energy increases due to data collection and processing overhead
Solution Approach 1:
The patent applies partial action by collecting and processing only the most relevant network usage data required for resource allocation decisions, rather than processing all possible network data. The learning algorithm focuses on extracting essential patterns from collected data, thereby maintaining high measurement precision for critical parameters while reducing overall energy consumption associated with comprehensive data processing
Data Source
AI summary
A software defined network controller for controlling a radio access network has a leaning capability by which usage characteristics are learned for each service provided by the radio access network. For example, bandwidth data, connection duration data and latency data may be used to learn usage characteristics for each radio access network service. The learned usage characteristics are then used in allocating network resources in response to a service request.