Dynamic Network Function Mapping to Hardware Resources
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for mapping software-based network functions to hardware resources in high-throughput network applications are inefficient due to inadequate consideration of microscopic and macroscopic design criteria, leading to suboptimal performance and increased data interchanges between processor cores and I/O buses, which results in reduced system throughput.
Innovation Solution
A dynamic method for mapping software-based network functions to hardware resources within a hardware resource pool, considering specific hardware constraints, network requirements, and detailed hardware descriptions, to optimize resource allocation and minimize data interchanges, thereby enhancing network performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If software-based network functions are distributed across multiple processor cores to meet high throughput demands, then processing capacity is improved, but data interchange between cores increases and becomes the bottleneck
Solution Approach 1:
The patent applies local quality by mapping specific processor cores to specific hardware resources (network interfaces, memory) based on the data interchange patterns required by each network function. This creates localized optimization where each core's working set is concentrated in nearby hardware, reducing the need for distant data interchange across the processor bus.
Solution Approach 2:
The patent implements dynamic mapping of software-based network functions to hardware resources based on runtime performance monitoring and traffic patterns. The mapping is not fixed but adapts dynamically to changing network conditions, allowing the system to optimize data interchange paths in real-time based on actual workload characteristics.
2Speed
If code is mapped to processing units to maximize cache hits and speed up communication, then communication speed is improved, but mapping complexity increases
Solution Approach 1:
The patent implements self-service through automated mapping mechanisms that monitor hardware resource usage and network traffic patterns. The system automatically adjusts code-to-hardware mappings based on observed performance metrics, eliminating the need for manual optimization while maintaining high communication speeds through data-driven decision-making.
Solution Approach 2:
The patent incorporates feedback loops that continuously monitor communication performance between processor cores and hardware resources. Based on this feedback, the system dynamically adjusts the mapping to optimize cache hit rates and communication efficiency, creating a closed-loop optimization system that adapts to changing conditions.
3Adaptability or versatility
If current grid computing and virtualization techniques are used, then resource sharing is improved, but mapping precision is inadequate for network software requirements
Solution Approach 1:
The patent applies segmentation by dividing the hardware resource pool into fine-grained units (individual cores, cache levels, memory banks) and mapping network functions to these segmented resources with precise control. This granular segmentation enables optimization at the hardware level while maintaining the virtualization benefits of resource sharing through software-based network function abstraction.
Data Source
AI summary
The method comprises mapping software-based network functions to hardware resources included in a hardware resource pool, wherein the mapping is performed dynamically on unallocated resources of the hardware resource pool and based on at least the next information:specific hardware constraints for the supporting of each of said software-based network functions;network requirements of at least one defined networking service design for a service making use of said hardware resource pool and of at least part of said software-based network functions; anda hardware description of the hardware resources included in the hardware resource pool.The system is configured for implementing the method of the invention.The computer program product is adapted to perform the method of the invention.