Improving Latency with Edge Computing and IoT in Health Applications: A Review (original) (raw)

World is developing at a quick pace as is data. The boom of smart phones and Internet of Things (IoT) enable more users to access data as well as computation power in real-time. The constraints of current distributed computing is a subject of more prevalent concern. In response to these rising difficulties in computing, as new requirements and challenges spring up day by day, especially in cloud computing scenario, computing paradigms that can meet these challenges are sought after. In order to expand the efficiency and to decrease the quantity of the data to be sent to the cloud for processing, numerous solutions have been proposed for the edge-centric network. In this paper, an elaborate study of computing architectures proposed for edge paradigms is carried out and application areas in the Internet of Things are ascertained. I. INTRODUCTION Computing architectures have gone through quite a few changes between centralized and decentralized models in the recent past. Cloud being the present centralized paradigm, majority of web content is served through a main data center and researchers rent their private servers from cloud for testing or experimenting. In addition Cloud offers an expedient way for both small and big businesses to obtain computing sources from a service provider instead of putting up their own data center. Usage of Cloud services like Google Cloud, AWS, Azure and VMware have become so popular in recent past. However, the new trend towards a soaring emphasis on edge devices and edge computing paradigms cannot go unnoticed. On the other side, many research organizations have given a startling prediction about the future computing scenario , that is, more than 20 billions of devices will be connected to the Internet by 2020. Thus more and more internet-enabled devices will result in generating huge amounts of data than ever. The Cisco white paper states that, IoT devices will generate 600 zettabytes of data by 2020[46]. Even though Cloud computing can be utilized in a pay-per-usage way through centrally managed resources, high latency and privacy issues are seen as solemn challenges. Hence, there is an inclination towards a decentralized solution of Cloud computing framework and the recent trend of computing paradigms lead to realization of shifting computing to the edge of the network. Since the connected devices prove to be potent in terms of computational capabilities and also battery power, they can be used for IoT applications. The requirements of low latencies and less bandwidth utilization of the data-intensive applications in the Internet of Things (IoT) are met. Most of the Cloud datacenters are centralized and located far from the nearness of the edge devices, and the latency-sensitive real-time service requests suffer due to large round-trip delays, degradation in service quality and network congestion. Such issues can be resolved efficiently by means of latest edge computing paradigms. Although the notions of edge and fog computing have been conceptualized before now, a universal understanding of these paradigms in practice is lacking. The rest of this paper is organized as follows: Section II highlights the earlier work found in literature. Section III discusses about the various edge paradigms based on their architectures. Section IV presents a comparison scenario of edge computing, fog computing and Mobile Cloud Computing (MCC) derived from an analysis of performance parameters. Section V identifies application areas for edge paradigms in the field of IoT and highlights unresolved deployment issues. Finally, Section VI concludes the paper.