Table 1 Comparison of related studies and the proposed EAVM protocol.
From: Energy-Aware adaptive virtualization and migration protocol for green IoT wireless sensor networks
Study/Year | Technique | Energy Harvesting | Learning Mechanism | Scalability | Sustainability Focus | Main Limitation |
|---|---|---|---|---|---|---|
Lightweight Virtualization Framework [25] | Static virtualization | ✗ | None | Low | Medium | No dynamic migration |
Predictive Migration Protocol [26] | Predictive analytics | ✗ | Centralized | Medium | Low | High migration overhead |
Federated Learning for Energy-Aware Task Allocation [28] | Distributed learning | ✗ | Federated Learning | High | Medium | No energy harvesting support |
Deep Reinforcement Learning for VM Placement [29] | VM allocation | ✗ | DRL | Medium | Medium | Lacks distributed scalability |
Federated DRL for IoT Traffic Offloading [30] | Resource offloading | ✗ | FDRL | High | Low | No energy harvesting integration |
Green IoT: Energy Efficiency, Renewable Integration, and Security Implications [31] | Green IoT framework | ✓ (renewable integration) | Policy-based | High | High | No virtualization and migration mechanism |
Energy-Aware Virtual Resource Scaling [34] | Adaptive virtualization | ✓ (limited) | Rule-based | Moderate | Medium | No FDRL integration |
Hybrid FL-DRL for Resource Allocation [39] | Hybrid federated + DRL coordination | Partial | Hybrid FL-DRL | High | Medium–High | Focuses only on aggregation efficiency, lacks virtualization/migration integration |
AoI-aware DRL for Energy-Harvesting IoT Networks [38] | Deep RL-based transmission control with adaptive energy management | ✓ (solar/RF) | Deep Reinforcement Learning (DRL) | Moderate | High | Focused on transmission optimization; lacks virtualization and migration mechanisms |
Proposed EAVM (this work) | Adaptive virtualization and migration | ✓ (hybrid solar/RF) | Federated Deep Reinforcement Learning (FDRL) | High (distribute) | High | — |
