AI-Driven MEC Service Allocation for Edge Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In Multi-Access Edge Computing (MEC) networks, the limited resources at edge locations can lead to reduced service availability and increased latency due to infrastructure constraints, as services may not be available at the requested edge location, forcing requests to be routed to the core network or external data network.
Innovation Solution
Implementing artificial intelligence and machine learning to dynamically allocate and manage services across edge locations based on changing usage patterns, prioritizing services that are in high demand, and reallocating resources to maximize service availability and reduce latency by shifting services to appropriate locations or networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If services are deployed at network edge locations to reduce latency, then service access speed is improved, but service availability is reduced due to limited infrastructure resources at edge locations
Solution Approach 1:
The system dynamically allocates and reconfigures service deployments across edge locations based on real-time usage patterns, demand predictions, and resource availability. Services are not statically assigned but continuously adjusted to optimize both availability and access speed, resolving the contradiction between limited edge resources and the need for high service availability.
Solution Approach 2:
The patent creates a universal service management system that can deploy any service across multiple edge locations and the core network. This multi-functional platform enables services to be flexibly located at the most appropriate position (edge or core) based on current conditions, thereby maintaining high availability while preserving low-latency access when needed.
2Reliability
If more services are deployed at edge locations to increase service availability, then service availability is improved, but resource constraints at edge locations are exceeded
Solution Approach 1:
The system segments services into different deployment categories based on their characteristics, usage patterns, and resource requirements. This segmentation allows the management platform to make informed decisions about which services should be deployed at edge locations versus the core network, optimizing resource utilization while maintaining service availability.
Solution Approach 2:
The patent changes key parameters such as service deployment location, resource allocation levels, and scaling factors based on analyzed usage patterns and predicted demand. By dynamically adjusting these parameters, the system can maintain high service availability without permanently over-provisioning edge location resources.
3Reliability
If services are cached at edge locations based on static usage patterns, then service availability is improved, but latency increases due to service thrashing when usage patterns change
Solution Approach 1:
The system implements continuous feedback loops that monitor actual service usage patterns, performance metrics, and resource utilization at edge locations. This feedback informs dynamic adjustments to service caching and deployment decisions, preventing service thrashing by adapting to changing usage patterns in real-time rather than relying on static predictions.
Solution Approach 2:
The patent uses machine learning models to predict future service usage patterns and proactively pre-positions services at appropriate edge locations before demand occurs. This preliminary action prevents service thrashing by ensuring services are already in place when needed, eliminating the latency associated with dynamic service migration during usage pattern changes.
Data Source
AI summary
Provided are systems and methods that use artificial intelligence and/or machine learning to dynamically allocate services at different times and at different network edge locations within a Multi-Access Edge (“MEC”) enhanced network based on a multitude of factors that change the priorities of the services at the different times and at the different edge locations. For instance, a MEC controller, controlling the allocation of resources at a particular edge location, may modify the allocation of services at that particular edge location at different times based on time and/or location sensitive events that occur at different times and that relate to different services, changing usage patterns that are derived from prior service utilization, and/or categorization of the services as permanent, time insensitive, or other categories of services with permissions to execute at different times from different edge locations.