1
Introduction to Aerial Aided Edge Computing: Understanding and Foundation of Edge Computing

Gifty Gupta1, Isha Kansal1, Charu Sharma2, Deepali Gupta1*, Kamali Gupta1 and Yonis Gulzar3

1Chitkara University Institute of Engineering and Technology, Chitkara University, Rajpura, Punjab, India

2M.M. Engineering College, M.M. (Deemed to be University), Mullana, India

3Department of Management Information Systems, College of Business Administration, King Faisal University, Al-Ahsa, Saudi Arabia

Abstract

Aerial-aided Edge Computing (AAEC) is a paradigm that incorporates the use of Unmanned Aerial Vehicles (UAVs) and edge computing to enhance information processing on the network edge. One major issue with the existing paradigms of cloud computing is that the passive nature of cloud services may be slow and could have low bandwidth especially in regions, where network access is unreliable or changes frequently. These challenges are solved for by edge computing that focuses on the data processing in regions closer to the data source thus enhancing the processing of networks, real-time data. This makes it possible to process the actual data and this in turns, shortens the time spent in working on the data, and also it minimizes the traffic on a particular network link. In addition to this, AAEC also utilizes aerial platforms for example drones which the authors also suggest to use as the mobile edge nodes which enhance the adaptability and scalability of the network. These aerial edge Nodes can be employed in places with poor or no grounding facilities like during disasters or in rural environment or in mobile networks. Besides this, AAEC goes a notch higher by employing flying modes to install mobile edge nodes which provide computing resources in areas that have scarce infrastructure. It leads to generation of more opportunities in various fields such as disaster management, smart city, and environment. This chapter presents the overview of edge computing, describes the key components of its architecture and unveils the possibilities of aerial platforms in enhancing the performance of edge computing for the first time. Additionally, it presents major technical concerns like power profile, reliable transfers and the problems of synchronizing the motion of the various edge nodes of the system so that a clear understanding of AAEC and its functions in future computing systems will be provided.

Keywords: Aerial-aided edge computing (AAEC), unmanned aerial vehicles (UAVs), mobile edge nodes, real-time data processing, disaster management and rural connectivity, network scalability and adaptability

1.1 Introduction to Edge Computing

Concept of edge computing has really advanced forward to solve the problems of latency and performance of the core cloud infrastructures. This has become necessary as the use of real-time application such as online virtual games, self-driving cars and applications for internet of things increases. Therefore, the latency requirements of these applications cannot be met by pure cloud approaches because of several factors that cause latency. The above viable issues are worked upon by Edge computing to provide services at the edge of the user devices [1]. Edge computing provides proximity benefit because it resides closer to the devices; hence, it minimizes the network delay and switches the data traffic direction, leading to substantial backhaul traffic and core network's traffic savings [2]. Microsoft, in the work of introducing mobile edge computing noted that the use of present day cloud architecture to satisfy the latency requirement for future service is inadequate due to the location of edge and the huge number of mobile phones. The application services can only succeed through the support, which the devices (vehicles, smartphones, drones, augmented and virtual reality devices) need to provide the proximity to the end-users and this is why the idea of the multi-access edge computing is considered to be essential. In addition, they described that local code execution in computation intensive devices can be achieved through code offloading from thin clients in mobile or static devices into the proposed mobile edge computing platform. As such, it improves the response time of the application and the services in an accurate manner to the end of the process. When Cisco explained the difference between centralized and decentralized edge computing, they argued that the latter can assist in achieving the targeted low latency response time, desired cost aspects, and data efficiency [3].

1.1.1 Definition and Concept of Edge Computing

Recently introduced, edge computing aligns with demands such as latency, throughput, privacy, and network stress generated by the future 5th and 6th generations of network technologies. However, the idea of edge computing is well-developed and promoted for several decades; still, various stakeholders define it according to their understanding and differently. That is why it is crucial to overcome the lack of a clear and integrated definition and the concept of edge computing. Therefore, in a literal sense, this paper defines edge computing as an architectural approach that relocates computation/processing functions closer to the data resource. In this literal sense, the main idea of edge computing is to utilize or create computing capabilities (that is, cloud) that are closer to the source of data than the inner cloud infrastructure [4].

When defined in a literal way, edge computing seems similar to the concept of cloudlet or mobile edge computing, which can be defined as a small-scale data center that provides powerful computational and storage capacities to mobile users typical to closely packed urban settings, even assuming functional differences (that is, cloud resources are assumed to serve resources designed uniquely for cloud computing, such as scalable processing power, storage, developing/testing platform, etc.). However, the definition of cloudlet and mobile edge computing is more precise that they concern with the resource service for special user terminal. Further, the steering of computing/processing steps closer to the data resources has been a practical postulate in systems design for a long time. Similar to the Internet architecture, there are DNS servers which need to be deployed closer to the edges since they respond to domain name resolution requests from end users who are immediately connected to the network and for reducing the latency of multimedia streaming, Content Delivery Networks (CDNs) introduce proxy/storing nodes in service access networks [5].

1.1.2 Evolution and Importance of Edge Computing

Edge computing has emerged due to the rising demand for faster and more efficient data processing as well as real-time analytics and appropriate low latency application. First, the scale of data processing was accomplished by cloud services, but with the development of IoT devices and high bandwidth applications, failures occurred in cloud computing. To decrease latency and utilize bandwidth more efficiently, edge computing appeared in order to process data on the edge of a network. Such technologies as 5G, AI, IoT contributed to this process. The essence of edge computing is to address the latency problem, security, bandwidth utilization to serve essential industries including healthcare, manufacturing, and autonomous systems which makes it crucial in the contemporary world [6].

1.1.2.1 Evolution of Edge Computing

Edge computing has emerged as an innovative concept because centralized cloud computing has proven to have various drawbacks due to advanced IoT request rates, other autonomous systems, and high bandwidth applications [6, 7]. Its journey can be understood in the following phases:

1. The Cloud Era (2000s - 2010s): The Cloud Era was characterized by the 2000s and the 2010s which saw most of the client applications offloading most of their computing and storage to large centralized data centers. But as IoT was growing, this model had its drawbacks and it could not work for tasks that needed low latency and response time. Cloud computing became very popular after many organizations moved their computing resources and data to servers in large remote data centers. This methodology proved very functional for many programs but caused problems in terms of waiting time and channel capacity for real-time as well as large data use cases. With a growing number of IoT devices, sending high volumes of data to the cloud for analysis was not effective as before. For instance, applications that need to support low latency that include autonomous vehicles, smart cities, and industrial automation could not afford to send data to a centralized cloud computing services since it would take time before receiving responses [8, 9].
2. Emergence of Fog Computing: Fog computing was brought up as a middle tier between the edge devices and cloud infrastructure. This approach sought to take cloud computing even further closer to the edge where some of the computations it entails are done in local servers or gateways. It paved way for creation more localized edge computing structures [10, 11].
3. Advent of Edge Computing (Late 2010s): In the late 2010s, the advent of Edge Computing brought change of model by decentralizing the data processing and networking to be more local to the source of data. In the subsequent year, developments in 5G, artificial intelligence (AI) and cutting-edge IoT devices went on to drive edge computing as...