Edge Computing Resource Allocation via QoE-Based Meta-Service Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current resource allocation methods in 5G networks, relying on cloud servers, fail to meet the requirements of low latency and high reliability due to latency overhead, necessitating a more efficient approach for fine-grained resource allocation based on user quality of experience (QoE) in edge computing.
Innovation Solution
Decompose application systems into meta-service units based on coupling degree, cluster them using QoE index parameter values, and allocate resources to each cluster according to these values, ensuring meta-service units with similar QoE parameters share resources, thereby enhancing QoE and resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If resources are allocated to each meta-service unit individually, then resource allocation precision is improved, but system complexity increases
Solution Approach 1:
The patent clusters multiple meta-service units with similar QoE characteristics into the same resource pool, allowing them to share resources collectively. This merging approach maintains fine-grained allocation precision for similar services while reducing the overall number of resource management entities, thereby lowering system complexity.
Solution Approach 2:
The patent introduces QoE index parameter values as the basis for clustering meta-service units. By changing the allocation granularity from individual units to clusters defined by QoE parameters, the system achieves both precise allocation for similar services and simplified management through parameter-based grouping.
2Device complexity
If cloud servers are used for resource allocation, then resource management is simplified, but latency increases
Solution Approach 1:
The patent segments the centralized cloud server resource management into distributed edge server resource pools. Each edge server manages its own resource pools locally, eliminating the need for centralized cloud server intervention in resource allocation decisions, thereby reducing latency while maintaining manageable complexity through automated clustering algorithms.
Solution Approach 2:
The patent introduces a resource pool as an intermediary layer between edge servers and meta-service units. This resource pool acts as a local resource management interface that abstracts the complexity of individual meta-service unit requirements, allowing edge servers to allocate resources efficiently without direct cloud server involvement.
3Ease of operation
If resources are allocated uniformly to all meta-service units, then allocation simplicity is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent applies different resource allocation strategies to different clusters based on their QoE characteristics. Each cluster receives resources tailored to its specific QoE requirements, allowing simple uniform allocation within clusters while achieving differentiated efficient allocation across the system through QoE-based clustering.
Data Source
AI summary
The resource allocation method and apparatus are based on edge computing. The method includes: decomposing all application systems in a server into a plurality of meta-service units according to the coupling degree between different service modules in the application systems, and obtaining the quality of experience (QoE) index parameter value of each meta-service unit; clustering the plurality of meta-service units by using a clustering algorithm based on the QoE index parameter value of each meta-service unit; and allocating resources to each cluster according to the QoE index parameter values of all meta-service units in each cluster, so that all meta-service units in each cluster share the allocated resources. According to the invention, the resource allocation is more reasonable, the QoE value of a user is increased, and the resource use is more effective.