MEC Cluster Resource Allocation via Dynamic Orchestrator
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
MEC clusters face challenges in efficiently allocating resources to prevent overloading and minimize latency, as existing systems struggle to dynamically manage CPU, memory, and bandwidth across devices, leading to potential service degradation when handling device mobility and varying computational demands.
Innovation Solution
The implementation of a resource allocation mechanism within MEC clusters that uses a cluster orchestrator to monitor resource utilization, predict future needs based on historical data, and strategically assign virtual network functions across MEC devices to balance load and ensure minimal latency, leveraging machine learning to optimize resource distribution between devices and a remote data center.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If resources are statically allocated to MEC devices, then device complexity is reduced and ease of operation is improved, but resource utilization efficiency deteriorates and service reliability worsens under varying computational demands
Solution Approach 1:
The patent implements dynamic resource allocation where the orchestrator continuously monitors resource utilization metrics (CPU, memory, bandwidth) and adjusts resource distribution in real-time based on current load conditions. This transforms the static resource allocation system into a dynamic one that adapts to varying computational demands, thereby maintaining service reliability without requiring complex manual intervention.
Solution Approach 2:
The system employs feedback mechanisms where the orchestrator receives utilization data from MEC devices, processes this information through machine learning models to predict future resource needs, and sends control decisions back to devices. This closed-loop feedback system ensures reliable service delivery by proactively adjusting resource allocation before performance degradation occurs.
2Reliability
If machine learning-based prediction is implemented to forecast resource utilization, then resource allocation accuracy is improved and service reliability is enhanced, but device complexity and computational overhead increase
Solution Approach 1:
The patent introduces an orchestrator as an intermediary component that centralizes the machine learning-based prediction and decision-making functions. Rather than embedding complex ML models in each MEC device, the orchestrator acts as a mediator that collects data from devices, performs sophisticated analysis using ML algorithms, and returns simplified allocation decisions. This approach enhances allocation accuracy while containing complexity in a centralized management entity.
3Reliability
If resources are over-allocated to prevent overloading, then service reliability is improved, but resource utilization efficiency deteriorates and waste increases
Solution Approach 1:
The system performs preliminary action by using machine learning models to predict future resource utilization needs before actual demand occurs. The orchestrator forecasts resource requirements based on historical patterns and current trends, then proactively allocates resources in advance. This prevents both overloading and over-allocation, as resources are reserved based on predicted needs rather than static over-provisioning, thereby avoiding waste while maintaining reliability.
Data Source
AI summary
A network device in a Multi-Access Edge Computing (MEC) cluster may receive a request for a service that requires use of a resource of a first type; and determine whether a resource utilization level, associated with the resource, at a MEC device in the MEC cluster exceeds a threshold. When the network device determines that the resource utilization level exceeds the threshold, the network device may determine whether a candidate MEC device is available within the MEC cluster to provide a resource of the first type; and enable the service to be provided.