Distributed AI Model Orchestration for Radio Access Network Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a need to introduce security for inference using AI models that need to be processed efficiently on constrained computing environments in radio access networks.
Innovation Solution
A method for executing workloads in a distributed system using an AI model, where the AI model is split into input blocks, an intermediate block, and output blocks, and the deployment configuration is defined based on current resource utilization status to optimize execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the AI model is processed on constrained computing environments in radio access networks, then the processing efficiency is improved, but security vulnerabilities increase
Solution Approach 1:
The AI model is divided into multiple blocks (first block, second block, third block) that are distributed across different computing systems. The first block remains in the constrained radio access network environment while the second and third blocks are placed in more secure cloud environments, thereby maintaining processing efficiency locally while mitigating security risks through spatial separation of model components.
Solution Approach 2:
An intermediary system (cloud computing system) is introduced to host the second and third blocks of the AI model. This intermediary provides enhanced security and computational resources, allowing the constrained radio access network system to benefit from secure model processing without requiring all model components to reside locally in the constrained environment.
2Object-affected harmful factors
If the AI model is split into multiple blocks and distributed across different systems, then security is improved, but system complexity increases
Solution Approach 1:
The computing systems are designed with multi-functional capabilities to handle different blocks of the AI model. The radio access network system can execute the first block locally while also coordinating with cloud systems for the second and third blocks, allowing a single distributed architecture to serve multiple security and processing requirements without requiring entirely separate specialized systems.
3Adaptability or versatility
If the deployment configuration adapts to current resource utilization status, then resource utilization efficiency is improved, but orchestration complexity increases
Solution Approach 1:
The deployment configuration is designed to be dynamic, automatically adjusting which computing system executes which block of the AI model based on real-time resource utilization status. When the radio access network system has sufficient resources, it can process more blocks locally; when resources are constrained, it offloads to cloud systems. This dynamic adaptation optimizes resource utilization without requiring manual reconfiguration.
Data Source
AI summary
The present disclosure relates to a method comprising receiving a request to execute a workload using an artificial intelligence model. A current resource utilization status in the distributed system may be determined. The current resource utilization status may be used to define a deployment configuration of the artificial intelligence model, wherein the deployment configuration is defined by: a number and structure of input blocks, a number and structure of output blocks and the intermediate block of the artificial intelligence model, a second computer system to execute the intermediate block, and one or more first computer systems to execute the input and output blocks. The artificial intelligence model may be deployed in accordance with the defined deployment configuration and the workload may be executed.