Fog Computing

60 60 Raza, U., Kulkarni, P., and Sooriyabandara, M. (2017). Low power wide area networks: an overview. IEEE Communication Surveys and Tutorials 19: 855–873.

61 61 Martin, B.A., Michaud, F., Banks, D. et al. (2017). Openfog security requirements and approaches. In: 2017 IEEE Fog World Congress (FWC), 1–6.

62 62 Sui, Y., Wang, X., Pengt, M., and An, N. (2017). Optimizing mobility and energy charging for mobile cloudlet. In: 2017 IEEE International Conference on Communications (ICC), 1–6. IEEE.

63 63 Tang, F., Fadlullah, Z.M., Mao, B. et al. (2018). On a novel adaptive UAV-mounted cloudlet-aided recommendation system for LBSNs. IEEE Transactions on Emerging Topics in Computing 7 (4): 565–577.

64 64 Truong-Huu, T., Tham, C.-K., and Niyato, D. (2014). To offload or to wait: An opportunistic offloading algorithm for parallel tasks in a mobile cloud. In: 2014 IEEE 6th International Conference on Cloud Computing Technology and Science (CloudCom), 182–189. IEEE.

65 65 Chang, C., Srirama, S.N., and Buyya, R. (2016). Mobile cloud business process management system for the Internet of Things: a survey. ACM Computing Surveys 49, pp. 70:1–70:42.

66 66 Li, M., Yu, F.R., Si, P. et al. (2018). Software-defined vehicular networks with caching and computing for delay-tolerant data traffic. In: 2018 IEEE International Conference on Communications (ICC), 1–6. IEEE.

67 67 Kumar, N., Rodrigues, J.J.P.C., Guizani, M. et al. (2018). Achieving energy efficiency and sustainability in edge/fog deployment. IEEE Communications Magazine 56: 20–21.

68 68 Ruan, Y., Durresi, A., and Uslu, S. (2018). Trust assessment for Internet of Things in multiaccess edge computing. In: 2018 IEEE 32nd International Conference on Advanced Information Networking and Applications (AINA), –1155, 1161. IEEE.

69 69 Wang, S., Xu, J., Zhang, N., and Liu, Y. (2018). A survey on service migration in mobile edge computing. IEEE Access 6: 23511–23528.

70 70 Zhang, P., Zhou, M., and Fortino, G. (2018). Security and trust issues in fog computing: a survey. Future Generation Computer Systems 88: 16–27.

71 71 Lipp, M., Schwarz, M., Gruss, D. et al. (2018). Meltdown: reading kernel memory from user space. In: 27th Security Symposium, 973–990.

72 72 Hussain, R., Son, J., Eun, H. et al. (2012). Rethinking vehicular communications: merging vanet with cloud computing. In: 2012 IEEE 4th International Conference on Cloud Computing Technology and Science (CloudCom), 606–609. IEEE.

73 73 Nobre, J.C., de Souza, A.M., Rosário, D. et al. (2019). Vehicular software-defined networking and fog computing: integration and design principles. Ad Hoc Networks 82: 172–181.

1 1 www.docker.com.

Cosmin Avasalcai, Ilir Murturi, and Schahram Dustdar

Distributed Systems Group, TU Wien, Vienna, Austria

In the past couple of years, the cloud computing paradigm was at the center of the Internet of Things' (IoT) ever-growing network, where companies can move their control and computing capabilities, and store collected data in a medium with almost unlimited resources [1]. It was and continues to be the best solution to deploy demanding computational applications with the main focus on processing vast amounts of data. Data are generated from geo-distributed IoT devices, such as sensors, smartphones, laptops, and vehicles, just to name a few. However, today this paradigm is facing growing challenges in meeting the demanding constraints of new IoT applications.

With the rapid adoption of IoT devices, new use cases have emerged to improve our daily lives. Some of these new use cases are the smart city, smart home, smart grid, and smart manufacturing with the power of changing industries (i.e. healthcare, oil and gas, automotive, etc.) by improving the working environment and optimizing workflow. Since most of the use cases consist of multiple applications that require fast response time (i.e. real-time or near real-time) and improved privacy, most of the time the cloud fails to fulfill these requirements (i.e. network congestion and ensuring privacy).

To overcome these shortcomings, researchers have proposed two new paradigms, fog computing and edge computing, to enable more computational resources (i.e. storage, networking, and processing) closer to the edge of the network. Fog computing (FC) extends cloud capabilities closer to the end devices, such that a cloud-to-things continuum is obtained that decreases latency and network congestion while enforcing privacy by processing the data near the user [2]. On the same note, the edge computing vision is to migrate some computational resources from the cloud to the heterogeneous devices placed at the edge of the network [3].

Embracing the vision of these paradigms and focusing on the deployment of multiple applications in close proximity of users, researchers have suggested new fog/edge devices. Among these devices, the most notable are mini servers, such as cloudlets [4], portable edge computers [5], and edge-cloud [6], which enable an application to work in harsh environments; mobile edge computing (MEC) [7] and mobile cloud computing [8] improve user experience and enable higher computational applications to be deployed on smartphones by offloading parts of the application on the device locally.

Many surveys are found in the literature that describe each paradigm in detail and its challenges [3, 9, 10]. However, there is no paper that compares the two; most of the time the terms fog and edge are both used to describe the same IoT network. Generally speaking, the visions of the two paradigms overlap, aiming to make available more computational resources at the edge of the network. Hence, the most significant difference is given by the naming convention used to describe them. The aim of this chapter is to offer a detailed description of the two aforementioned paradigms, discussing their differences and similarities. Furthermore, we discuss their future challenges and argue if the different naming convention is still required.

The remainder of the chapter is structured as follows: Section 2.2defines the edge computing paradigm by describing its architectural features. Next, Section 2.3presents in detail the fog computing paradigm and describes two use cases by emphasizing the key features of this architecture. Section 2.4describes several illustrative use cases for both edge and fog computing. Section 2.5discusses the challenges that these paradigms must conquer to be fully adopted in our society. Finally, Section 2.6presents our final remarks on the comparison between fog and edge computing.

As we explore new IoT applications and use cases, the consideration of proximity between edge nodes and the end-users is becoming increasingly obvious. The physical distance between the edge and the user affects highly end-to-end latency, privacy, network, and availability. Recently, this leads to a new paradigm allowing computation to be performed in close proximity of user and IoT devices (i.e. sensors and actuators). Edge computing [11] is a new paradigm aiming to provide storage and computing resources and act as an additional layer, composed of edge devices, between the end-user IoT device and the cloud layer. In edge computing, we define “edge” as any computing and network resources along the path between the initial source of data and destination storage of data (fog nodes, cloud data centers).

Figure 2.1 Edge computing solution using an IoT and edge devices [12].