Network Function Arrangement Device Using Pre-computed Resource Statistics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In existing systems, implementing network functions across multiple types of hardware requires lengthy simulation and operation checking, leading to resource wastage due to differing processing performance and resource consumption across hardware types.
Innovation Solution
An arrangement device that includes a storage unit for statistical information on resource consumption and performance, an accepting unit for high-level language descriptions, a prediction unit to calculate predicted performance and resource consumption, and a selection unit to choose hardware that meets performance requirements and resource capacity, allowing quick arrangement of network functions across various hardware types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If simulation and operation checking are performed for each hardware type to verify processing performance and resource consumption, then the reliability of the arrangement decision is improved, but the time required for arrangement increases significantly
Solution Approach 1:
The system performs simulation and operation checking in advance for multiple hardware types, storing the results as statistical information including processing performance and resource consumption. When a network function needs to be arranged, the pre-computed statistical information is retrieved and used for quick comparison and decision-making, eliminating the need to perform simulations at the time of arrangement.
Solution Approach 2:
The system pre-calculates and stores statistical information about processing performance and resource consumption for various hardware types and network function combinations. This beforehand preparation creates a cushion of ready-to-use data that enables rapid arrangement decisions without compromising reliability, as the pre-computed data can be directly applied to new arrangement scenarios.
2Loss of energy
If multiple network functions are arranged in the accelerator to improve resource utilization, then resource efficiency is improved, but the complexity of arrangement decision increases
Solution Approach 1:
The system pre-calculates resource consumption statistics for multiple network functions on different hardware types and stores them. When arranging functions, the decision unit retrieves these pre-computed statistics and automatically selects hardware that optimizes resource utilization, enabling efficient multi-function arrangement without manual complexity.
Solution Approach 2:
The arrangement decision unit automatically selects appropriate hardware and determines optimal function placement based on the stored statistical information, without requiring manual intervention. The system self-services the complex arrangement decision by using the pre-computed data to automatically optimize resource allocation across multiple functions and hardware types.
Data Source
AI summary
A storage unit stores statistical information including an amount of resource consumption and performance information, which represents a performance, of each piece of hardware of a plurality of types that are candidates for an arrangement destination of a function, an accepting unit accepts inputs of description details of a function in a high-level language corresponding to the hardware of the plurality of types, and a performance requirement that represents a required performance, a performance predicting unit calculates a predicted performance, and a predicted amount of resource consumption, using the description details and a predetermined algorithm for each piece of hardware; and a device selecting unit selects, as an arrangement destination, hardware with the calculated predicted performance and the performance information satisfying the performance requirement and a total value of the predicted amount of resource consumption and the amount of resource consumption being equal to or smaller than a resource capacity.