The wireless network paradigms are growing rapidly; the last decade is the witness of it. Networks' capabilities flourished from third generations (3G) to fifth generation & beyond (5GB) steps with the expatriation of high node mobility. Therefore the mobility management became an integral part of wireless networks. The coupling of wireless networks and mobility management is one of the most essential areas of research and development where it can shift-up to the next level for better quality of service (QoS) and quality of experience (QoE).
To achieve the user's experience in wireless networks, it is associated with cloud computing and the internet of things (loT) and gradually it moves towards fog and edge computing-based networks. It provides a dynamic association and enhanced mobility management procedure. People are looking forward to adding machine learning (ML) system models for intelligent network behaviors. Traditionally, local nodes collect the required data and process the data and transfer it to the remote cloud and system train machine learning models, with the results. The next possibilities are co-located local devices perform the required training, so local data never needs to be uploaded. Using it we can improve the QoE. Nowadays loT plays a vital role in modern wireless networks. Therefore, loT is shifted to the Internet of Everythings (loE). This predominance generates many loT based applications such as Healthcare-technology-enabled loT networks. This infrastructure consists of many legacy medical sensors, loT-based personal health devices, and software applications. It generates a huge amount of medical data that need to be processed, correlated and analyzed in near real-time computing. Cloud computing and is used for computing in big data. Unique mobility management is mandatory to serve the dynamic schema where an accurate precision and reliability is needed. Such massive data collection and aggregating with appropriate device/user/system require an effective cloud-based data processing system and computation techniques to secure, as well as reliable integration of enterprise loT networks with public and private cloud networks, as well as personal-connected health devices. The detailed analysis is thoroughly discussed in this book. The healthcare loT is very often associated with body area networks (BAN). This technology consists of different sensor nodes implanted in the human body to read and analyze the physiological health information (PHI) parameters such as blood sugar, blood pressure, body temperature, etc. If any PHI reading is beyond the normal range of the corresponding PHI, the medical event is reported instantly. These types of works are in the scope of the book.
The wireless technology is not limited to a small geographical area. It is gaining attention for wide areas such as smart cities and smart homes. It also propels us towards several challenges. The loT can be a good bribe between smart cities and smart wireless networks. The technical amalgamations are equipped with sensors, microcontrollers, transceivers, provided the unique identity (RFID, Barcode, etc.), wireless connectivity and a suitable protocol stack for a seamless connection over a data network. This book will provide a framework to provide a comprehensive study for smart cities, loT and cloud-based computing models. But it also carries several research challenges and cutting edge technological improvements including security, scalability, research and methodology and cost- effectiveness that are well elaborated in this book.
As the data is a stored and processed over the cloud, therefore it enables several risks and security aspects. Cloud-based loT is the one approach to overcome it. The distributed nature of a Cloud-based loT infrastructure is prone to different threats and vulnerabilities related to technological and human-centric factors as well as strategic decisions in the design and implementation of a Cloud-based loT. This book deals with all of these solutions and methods.
The baseline of cloud networks is a TCP/IP reference model. The model is old enough, and it leads us to new networks theories such as software-defined networks, information-centric networks (ICN), fog and edge networks. The book helps us to find the limitations of current Internet architecture, data processing, information transmission and highlighting the importance of ICN, as it is used to overcome the limitations of legacy network architecture. The significance of ICN with loT over the cloud is also well explained in this book.
I: Evolution of IoT, Cloud Network and Network Mobility
Evolution of Cloud Fog IoT Interconnection Networks
Department ofCSE, National Institute of Technology, Rourkela, Odisha
The traditional cloud computing architecture built on virtual machine (VM) based offerings is faced with multiple challenges such as failure to provide performance guarantee for bandwidth and propagation delays. This made the platform undesirable for modern applications that are network-intensive and have stringent requirements on tolerable delays. To conquer this issue, infrastructure providers (InPs) started to offer resources in the form of virtual data centres (VDCs), that have multiple interactive VMs distributed geographically with different performance requirements. Although services offered through VDCs brought about benefits in terms of improved services and better user experience, issues such as transmission delays and communication overheads were still persistent that made the platform undesirable for the class of latency sensitive applications. As a solution, fog computing was introduced, which brought cloud services closer to the users and thereby reduced the dependency on centralized cloud. Moreover, with the rapid proliferation of Internet of Things (loT) devices, servicing latency sensitive applications with cloud at the backbone was not feasible. In order to cater to the requirements of such complex applications an interplay between the three stratum, i.e., loT, Fog and Cloud, is essential. In this chapter, we discuss the evolution of cloud networks starting from the traditional VM based offerings to a more complex model that involves a fruitful interplay between the stratum. Every stage of evolution posed different challenges of which some were extensively researched and of some that did not receive much attention. We study in detail the issues and discuss possible solutions proposed in the literature to address them. Further, we also discuss the strengths and weaknesses of each solution and highlight areas where open research can be conducted.
The traditional virtualization technology was mainly focused at delivering services in the form of computing and storage resources. Although it was able to overcome the drawbacks of traditional data centres (DCs), such as low resource utilization, higher operational costs and lack of isolation, it failed to provide performance guarantee for bandwidth and propagation delays across DC networks. With the outburst of cloud computing, it emerged as a desirable platform for network-intensive applications such as Hadoop . The performance of such applications is heavily reliant not only on computing and storage resources but also on networking resources on which they are deployed. Hence, traditional VM based services were not adequate enough to meet the desired quality of services (QoS) for network-intensive applications. To overcome such limitations, infrastructure providers (InPs) started to offer resources in the form of virtual data centres (VDCs). A typical VDC request is depicted in Figure 1.1. As can be observed from Figure 1.1, a typical VDC request comprises multiple communicating VMs with performance guaranties on delay experienced. Each component of a VDC, i.e., VM or VL, has different resource demands. Coming back to Figure 1.1, the values corresponding to a VM consist of vCPU, memory and disk image demands, respectively. The values corresponding to the VLs denote link resource demands and maximum tolerable transmission delay, respectively. From an InPs perspective allocation of resources in the form of VDCs is posed with a variety of challenges. In the next section, we discuss in detail such issues and solutions proposed in the literature to tackle them.
Although partitioning physical resources at the DCs into VDCs brought about benefits in terms of improved service guaranties and better user experience, issues such as transmission delays and communication overheads were still persistent . In fact cloud computing was still not a viable platform to service cloud based latency sensitive applications such as connected vehicles, fire detection, fire fighting, smart grid and content delivery . To address this issue fog computing was proposed that extends the services of cloud closer to the users. Specifically, fog computing enables processing of latency-sensitive, real-time and
Figure 1.1: A typical virtual data center request.
responsive applications at the network edge whereas latency-tolerant applications can be processed at a distant cloud. In fact fog computing is pivotal, particularly when it comes to services supporting data management and analytics.
The rapid growth of Internet of Things (loT) devices is fuelled by the need for smart objects, including sensors, smart meters, smart cars and actuators, to collect and exchange data to facilitate various applications, such as smart city, smart grid, e-healthcare and home automation . These devices generate huge amounts of data and require almost instantaneous processing, mobility support, geo-distribution in addition to location awareness and low latency that make the traditional cloud an inappropriate platform for such applications. Alternatively, fog computing due to its inherent characteristics can be used as a platform for executing loT applications . Hence, it can be perceived that rather than cannibalizing cloud computing, fog computing is an enabler to a new breed of applications and services that thrives on a fruitful interplay between the stratum.
Every stage of evolution starting from the traditional VM based offering to VDCs and then to an interplay between fog-cloud and loT is posed with different challenges of which some were extensively researched and of some that did not receive much attention. We study in detail such issues and discuss possible solutions proposed in the literature to address them. Further, we also discuss the strengths and weaknesses of each solution and highlight areas where open research can be conducted.
The rest of the chapter is organized as follows. The motivation and contributions are highlighted in Section 1.2. Section 1.3 deals with the evolution of traditional VM-based offering to virtual data centres (VDCs) and the challenges associated with it. Section 1.4 discusses the need for fog computing and identifies the challenges associated with it. Section 1.5 discusses the challenges and applications of a more complicated cooperation between loT-Fog-Cloud. Finally, some conclusions are drawn in Section 1.7.