A survey of mobility-aware Multi-access Edge Computing: Challenges, use cases and future directions

Many mobile and pervasive applications avail cloud services to reduce overheads in on-device computation. The performance of these services depends on the available bandwidth of the underlying network, the physical proximity of the cloud server and the end devices, the volume of data, the computational capacity of the server, and, importantly, the mobility of the devices hosting the applications. Edge computing promises to provide better performance by bringing services (e.g., a video streaming service) from the cloud to servers near the user. It also enables partial or full offloading of the computation (tasks) and storage functionalities from the User Equipment (UE) to the edge of the network. This saves power and benefits from relatively more powerful devices at the edge. Multi-access Edge Computing (MEC), which supports wireless and wired access technologies, has gained significant research interest. When UEs move, services must continue to operate, tasks may need to be offloaded again, and states related to tasks and services may need to be migrated. In this paper, we focus on four functional components (task/service offloading, resource allocation, content/task caching, and service/task migration) of MEC. We survey the challenges to these and their solutions in the context of UE mobility. Mobility creates challenges during offloading resource-intensive tasks as the user may move while the task is being offloaded. Some of the other challenges are how to jointly allocate computing and communication resources, minimize service down time during migration, and share the backhaul network if the same MEC host must continue to be used. Some key research areas include intelligent task offloading and service migration algorithms, exploiting group mobility to improve task migration time, studying the interplay of MEC parameters such as capabilities of the target MEC host, etc. In addition, predicting the mobile trajectory through intelligent methods and implementations with datasets from real-world scenarios are required.We compare this paper on 11 parameters (service migration, task offloading, resource allocation, content caching, mobility, use cases, architecture, computing paradigm, mobility model, system model, virtualiza-tion/Software Defined Networks) with 31 other survey papers from 2018 to April 2022 in MEC and related domains. We discuss the Edge Computing paradigm, the system architecture and model descriptions, and use cases. We briefly explain the relevant challenges and future directions in emerging domains, such as the Internet of drones and Digital twins. We also discuss future research directions in task/service migration, offloading, resource management, distributed computing, reliability, and Quality of Service, all related to mobility in MEC.