Introduction
As we transition into the era of 5G and 6G, the deployment of ultra-dense networks has become a necessity to support the massive influx of data and low-latency requirements. Multi-access Edge Computing (MEC) is a cornerstone of this evolution, bringing processing power closer to end-users to facilitate applications like autonomous driving and augmented reality. However, ultra-dense environments introduce a significant challenge: the "digital handoff."
In these networks, base stations are packed so tightly that high-mobility users—such as those in fast-moving vehicles or high-speed trains—cross cell boundaries every few seconds. Each crossing necessitates a service migration, where the user's digital state must move from one edge server to the next. Frequent migrations can saturate backhaul networks, leading to service interruptions and suboptimal server selection based on raw metrics that ignore actual human perception.
Hierarchical MEC Architecture
To address these challenges, this research proposes a hierarchical MEC architecture designed to streamline decision-making and reduce signaling overhead. By shifting the handover logic to the MEC server level rather than the core network, the system can achieve significantly lower latency.
Technical Innovations: CMVAM and MEEQ
The framework introduces two primary mechanisms to optimize the migration process:
1. CMVAM (Chain Management with Valuation Adjustment Mechanism)
Borrowing principles from financial options, CMVAM treats network resources as dynamic assets. Users are assigned a VAM code, representing a credit level. If a user migrates and releases resources responsibly according to their service contract, their credit rises; conversely, if they cause network strain, it drops. This credit-based system allows the network to prioritize the migration queue during congestion, ensuring that high-priority, high-mobility services are restored first.
2. MEEQ (Migration Effect Evaluation with QoS-aware)
Traditional Quality of Service (QoS) metrics often fail to capture the non-linear nature of human perception. MEEQ utilizes a Sigmoid-based mathematical model to divide performance into three distinct zones:
- Best: Where additional bandwidth provides no perceptible benefit.
- Sensitive: Where small performance changes are highly noticeable to the user.
- Unusable: Where the service fails to meet basic requirements.
By focusing on keeping users out of the "Sensitive" and "Unusable" zones, MEEQ allows for more resource-efficient management of crowded edge environments.
Experimental Evaluation
The researchers validated their framework using a custom simulation environment built with the networkx Python package. The simulation featured a hexagonal grid of 49 base stations, each equipped with an MEC server. The environment tested a mix of "ordinary" users (staying in cells for 5–30 minutes) and "high-mobility" users (staying for only 5–50 seconds).
Key Results
- Enhanced Restoration: In network "blockage" scenarios, CMVAM restored services significantly faster than the standard Follow-Me Chain (FMC) algorithm, particularly when high-mobility traffic was dominant.
- Resource Efficiency: The advantages of CMVAM were most evident when Virtual Network Function (VNF) capacity reached 40%, proving its efficacy in resource-constrained, dense environments.
- Perceptual Accuracy: MEEQ’s satisfaction scores provided a more accurate reflection of the actual human experience of service degradation compared to traditional linear assessment methods.
Conclusion and Future Work
By combining credit-based prioritization with human-centric QoS evaluation, this framework offers a robust solution for service continuity in the next generation of mobile networks. While the current simulation results are promising, future research will focus on integrating these mechanisms with Software Defined Networking (SDN) and testing against more complex, real-world mobility patterns and physical layer interference.
Summary updated