Introduction
With the rapid advancement of Mobile Edge Computing (MEC), a vast array of services from diverse domains has become accessible at the network's edge. These services often collaborate to form complex networks, providing users with rich resources and low-latency responses. However, the sheer volume of available services and the dynamic nature of Quality of Service (QoS) present a significant challenge: discovery overload. Users struggle to identify "premium" services—those that are highly reliable and effective—amidst a sea of options.
Traditional network models, which rely on simple graphs with nodes and pairwise edges, fail to capture the multi-party interdependencies common in modern service ecosystems (e.g., a video streaming service collaborating simultaneously with transcoders and CDNs). To address these challenges, this research proposes a novel method: Perception of Premium Services based on the evolution of Hyper-Service Network (PPSHSN).
Modeling Multi-Service Collaboration
The core innovation of PPSHSN lies in the use of hypergraph theory. Unlike standard graphs, a hyperedge in a hypergraph can connect any number of nodes, making it an ideal mathematical structure for representing multi-service collaborations. By modeling the service ecosystem as an evolving hypergraph, the researchers can track how services join, collaborate, and leave the network over time.
The PPSHSN Methodology
The PPSHSN framework consists of three primary phases designed to proactively perceive the emergence of top-tier services:
- Hypergraph Construction: The network is built using a "linked list method" for preferential connection. In this model, new services are more likely to connect to existing services that already exhibit high usage frequencies, mimicking real-world popularity trends.
- Evolution Mechanism: The network is treated as a living system. Rules are established for adding new nodes and hyperedges based on cooperation strength, while idle or underperforming services are pruned to maintain efficiency.
- The HDPV Metric: To identify premium services, the authors developed the Hyperdegree Premium Value (HDPV). This metric integrates a service's topological importance (Node Hyper-strength) with its actual operational reliability (Success Rate).
Evaluation and Key Results
The effectiveness of PPSHSN was verified through extensive simulations using a hybrid dataset. This dataset combined real-world context—derived from a survey on online video processing services—with synthetic QoS attributes such as availability, reliability, and reputation.
Scalability and Efficiency
A critical finding was the efficiency of the "linked list" construction method compared to traditional "roulette wheel" selection. In tests involving 2,000 nodes, the linked list method completed construction in 11.55 seconds, whereas the roulette wheel method required 31.79 seconds. This scalability is vital for real-time MEC environments where service availability changes rapidly.
Stability and Accuracy
The HDPV metric proved to be more stable and accurate than baseline topological metrics. While simple node hyperdegree can fluctuate wildly as the network evolves, HDPV consistently identified the same premium services. The evolution process naturally filtered the ecosystem, allowing highly reliable "hub" services to emerge as dominant nodes following a power-law distribution.
Conclusion and Future Work
The PPSHSN method provides a robust framework for identifying high-quality services in complex, dynamic MEC environments. By leveraging hypergraph evolution, the model captures the nuances of service "teamwork" that traditional models overlook.
While the current study focuses on single-layer networks, future research will aim to extend this evolution model to multi-layer architectures, reflecting the physical and virtualized layers of modern cloud-edge infrastructure. Additionally, further refinement of weighting parameters will ensure the discovery process remains optimized for even larger, industrial-scale IoT datasets.
Summary updated