As the metaverse progresses, it requires real-time, low-latency, and secure computing capabilities that conventional cloud-based systems cannot fully support. Multiaccess edge computing (MEC) addresses this demand by moving processing closer to the user; however, it also introduces new challenges in task placement, resource management, security, and trust. The blockchain is a promising enabler to address the limitations of trust, transparency, and centralized control in these systems. This survey systematically reviews 61 high-quality studies that explore blockchain-based solutions for task offloading and resource allocation in MEC. This work identifies the prevailing trends and research gaps using a structured method. The analysis reveals that over 60% of studies employ artificial intelligence-based techniques. Additionally, this work quantifies the adoption of various performance metrics from the literature and examines the case study distribution across application domains. Finally, this work outlines the technical challenges and opportunities for future research toward scalable, efficient, and trustworthy blockchain-enabled MEC frameworks.

Delivering realistic, real-time virtual experiences in the Metaverse demands computationally intensive rendering tasks with strict latency constraints. These challenges are further intensified by dynamic behavior and heterogeneity of network resources. While traditional machine learning (ML) methods have been applied to task placement, they often fail to accurately capture the complex relationships between tasks and computing nodes in such dynamic environments. To overcome these limitations, this study introduces a graph neural network (GNN)- based adaptive framework for in-network placement of delay-sensitive tasks in the Metaverse, leveraging the computing-in-the-network (COIN) paradigm. The task placement challenge is modeled as an integer linear programming (ILP) problem aimed at minimizing overall system costs while adhering to stringent delay constraints. Although optimal solutions can be derived using standard optimization solvers, their high computational overhead makes them unsuitable for real-time deployment. To address this, the optimal ILP solutions train a GNN offline, enabling fast, efficient placement decisions at runtime. The proposed GNN framework captures connected and isolated node interactions, facilitating dynamic adaptation to network conditions. It achieves a placement accuracy of 92.44%, surpassing multilayer perceptron 76.25% and decision trees 70.16%. Simulation results further confirm that the proposed model substantially outperforms conventional ML approaches, while task partitioning across multiple COIN nodes enhances network resource utilization.