Resource Allocation via Consumption Forecast in MEC
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In MEC (Mobile Edge Computing) infrastructure, decentralized data centers face resource constraints, leading to potential shortages when multiple client entities simultaneously request resource allocation for processing services, disrupting network services like electricity consumption reading applications.
Innovation Solution
A method where the application service manager entity notifies the network services manager entity of a consumption forecast, allowing for cooperation and anticipation of resource needs, enabling efficient allocation and avoiding resource monopolization by a single entity, with features like intermediate consumption levels and resource reservation to optimize service quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If decentralized data centers are deployed closer to access points to reduce latency, then response time is improved, but resource availability deteriorates
Solution Approach 1:
The application service manager entity sends a consumption forecast notification to the network services manager entity in advance, before the actual resource allocation request. This preliminary action allows the network services manager to anticipate future resource needs and prepare accordingly, ensuring resources are available when needed without requiring excessive resources to be constantly maintained at decentralized data centers.
Solution Approach 2:
The system establishes a feedback loop where the application service manager communicates consumption forecasts to the network services manager, who then adjusts resource allocation based on this information. This feedback mechanism enables dynamic resource management that responds to actual service needs while optimizing resource utilization across the distributed infrastructure.
2Adaptability or versatility
If multiple client entities simultaneously request resource allocation, then service coverage is improved, but resource shortage occurs
Solution Approach 1:
By requiring consumption forecast notifications before resource allocation requests, the system allows the network services manager to anticipate and prepare resources for multiple simultaneous clients. This preliminary planning prevents resource shortages even when multiple entities request resources at the same time, as the manager can allocate resources based on forecasted needs rather than reacting to concurrent requests.
Solution Approach 2:
The application service manager entity is responsible for generating and sending its own consumption forecast to the network services manager. This self-service approach allows each client entity to actively participate in resource planning, enabling the network services manager to better coordinate resource allocation across multiple simultaneous requests and prevent conflicts.
3Productivity
If execution resources are allocated to application services, then service functionality is improved, but network service quality deteriorates
Solution Approach 1:
The network services manager entity receives consumption forecast notifications in advance and can prioritize resource allocation to ensure network services maintain sufficient resources. By knowing future application service resource needs beforehand, the manager can pre-allocate or reserve resources for critical network services, preventing quality degradation when application services are deployed.
Data Source
AI summary
A method of allocating execution resources, by a virtualized-resources manager entity, for an execution of an application service and of at least one network service. The execution of the application service depends on the concurrent execution of the at least one network service. the method includes: a first request to allocate execution resources by a manager entity of the at least one network service to the virtualized-resources manager entity; a second request to allocate execution resources by a manager entity of the application service to the virtualized-resources manager entity; and, prior to the requests, a notification, by the manager entity of the application service, of consumption forecast by the application service of at least one network service provided by the network services manager entity, to the services-managing network entity.