1. What are the challenges and open issues in edge computing optimization methods?
Edge computing optimization methods face several challenges and open issues. These include the need to consider all metrics, architecture, and optimization methods for a comprehensive understanding of the field. The existing surveys in the literature review have highlighted the importance of presenting a work that provides a complete overview of optimization techniques, including all metrics and details. The aim of this paper is to merge the three layers of architecture with optimization methods and include computation offloading of tasks in edge computing. The challenges and open issues in edge computing optimization methods are discussed in Section III of the paper. The main and possible future research directions are also discussed in Section IV. Overall, addressing these challenges and open issues is crucial for advancing the field of edge computing optimization methods.
read more
2. What factors affect task offloading decision?
Factors affecting task offloading decision include optimization methods, energy consumption, elapsed time, and processing location. Optimization methods involve machine learning, heuristic algorithms, and scheduling algorithms. Energy consumption and elapsed time are calculated for each task in terms of processing and transmission. Processing location refers to the location of task processing, such as IoT device, MEC server, or the cloud. These factors are crucial in determining the most efficient offloading decision to improve system performance.
read more
3. What is the hierarchical architecture in edge computing?
The hierarchical architecture in edge computing consists of the cloud layer, edge layer, and IoT layer. This architecture is presented in Figure 2 [20] and considers the user layer identified by wireless communication used by devices [21]. The fundamental distinction between the edge and cloud layers is the computing power provided by servers at the edge and in the cloud, respectively. Sensors, mobile phones, and automobiles are among the devices in the user layer [22]. Computation-intensive operations are transmitted from these devices to distributed edge servers for processing [23].
read more
4. What are the key considerations in optimizing energy consumption and latency in edge computing?
Several studies have presented the effect and importance of energy and latency in edge computing. Some researchers have focused on IoT devices and MEC, while others have moved functionality from the cloud server to the edge server. Different approaches have been proposed, such as online reinforcement learning, distributed offloading, deep Q-network-based strategic computation offloading, transmit power allocation, energy-efficient computation offloading, Joint Caching and Wireless and Backhaul Scheduling (JCWBS) algorithm, fast numerical algorithms, data placement optimization, deep reinforcement learning-based joint optimization, online Double Deep Q Networks (DDQN) based learning scheme, enumeration-Based Optimal Edge Server Placement Algorithm (EOESPA), Ranking-based Near-optimal Edge Server Placement Algorithm (RNOESPA), alternating optimization algorithm, binary-coded genetic algorithm, constrained multi objective evolutionary algorithms (CMOEAs), joint iterative algorithm, and Colony Optimization algorithm. However, most of these studies do not consider the cloud layer, which is an important aspect of edge computing. Therefore, it is crucial to consider the cloud layer when optimizing energy consumption and latency in edge computing to achieve a comprehensive and effective solution.
read more