With the proliferation of Internet of things which potentially connects billions of devices, there is tremendous increase in generation and flow of data, the traditional cloud system used to carry out data analytics and decision-making process in certain cases can encounter a series of challenges particularly due to centralized processing, storage, network, bandwidth and latency issues. In order to tackle these problems a new paradigm known as edge computing seems to be a promising technology to revolutionize traditional IoT ecosystem. It aims at shifting the processing capabilities of centralized cloud computing to the edge of the network devices thus reducing the burden of the centralized cloud data centers. The applications of edge computing can be found in various domains such as e-commerce, transportation, environment and healthcare. In this paper, definition and architecture of edge computing, its application in healthcare sector and certain challenges and opportunities are summarized.
We are currently living in the era of Internet of things which in general terms involves interconnection of smart objects with sensors, actuators, software and network connectivity. Internet of things enables communication of these edge devices with each other as well as with human beings. Data is progressively produced at the edge of the network and therefore its processing at the edge would prove to be more efficient. According to a report from Cisco , “The ‘Internet of Everything’ — the people and things connected to the internet — will generate 507.5 zettabytes (1 zettabyte = 1 trillion gigabytes) of data by 2019”. Currently, billions of devices are connected by internet of things and this number is continuously increasing thus creating unaccustomed and variety of data. There are certain challenges faced by Internet of things such as scalability, interoperability, massive data, high latency, heavy bandwidth utilization, security and privacy etc. which are limiting its applications. In order to combat these challenges a new paradigm called edge computing is seen as a promising solution which can lay the foundation for the evolution of new age applications and services.
Edge computing permits the data produced by internet of things (IoT) devices to be processed closer to where it is generated instead of sending it across long paths to cloud or data centers.
Edge computing can be defined as a network of micro data centers that store or process relevant data locally while shoving all the received data to a cloud storage unit, in propinquity of the source of data produced.
Edge computing involves decentralized processing and is a method of optimizing cloud computing by performing data processing near the source of the data i.e. at the edge of the network. This is particularly beneficial as cloud computing is not always efficient and has certain limitations. The prime motive behind edge computing is that the processing of data should occur at the proximity of data sources. Take the case of internet of things where the edge devices collect enormous data and send it to the central data center or cloud for its processing edge computing triages the data locally thus reducing flow of data to the central repository. Typically, the whole process involves transferring data from IoT devices to a local device present in vicinity that has the ability to compute and store and has embedded software and network connectivity. Processing of data takes place at the edge and a part of it is sent to the processing or storage repository in a data center, co-localization facility or cloud.
Apart from Internet of things, edge computing can also be blended with mobile ad-hoc networks (MANETs), vehicular ad-hoc networks (VANETs), and intelligent transport systems (ITSs).(transational ieee)
Edge computing adds an additional layer of computing power downstream of cloud data centre and upstream of edge devices, keeping critical data closer to the source device thereby reducing the processing and response time. The individual devices themselves become junctions that are capable of controlling and maintaining relatively smaller time-sensitive tasks without the need of having it sent to the cloud.
(few meters away)
Gateway and Network
Edge computing benefits
Data is processed close to the source of data thus reducing latency between edge device and the processing unit and enabling faster decision making. This is particularly helpful in time critical data.
The data is localized and if any of the device malfunctions, it does not affect the result of the other devices as it would have done in the case of cloud computing. As the data stays locally, only the insights flow so there is less chance of violation of any compliance or regulation. Preserving the data locally also reduces the chance of all data getting attacked at once by hackers thereby increasing the security. The computation of data at the edge also helps in reducing response time to a few milliseconds thus making real-time decisions required in certain tasks.
Table 1. Comparison between cloud and edge computing(El-sayed, )
High bandwidth utilization
Very low bandwidth utilization
Response is in few minutes or even weeks
Response is in few milliseconds
Energy consumption is high
Low energy consumption
Could take time in decision making
Takes near real-time decisions
Edge computing in healthcare
The proliferation of internet of things is revolutionizing healthcare sector by speeding up diagnostic and treatment processes, enhancing quality of treatment, enabling remote monitoring of patients, clinical trials and reducing return visits therefore cutting entities money. By the end of 2020, the total number of smart healthcare devices is expected to reach 808.9 million ( 646 million devices without wearables and the rest with wearables).()() The amount of data generated from connected internet of things, medical monitoring systems, mobiles and various other sources is continuously growing which increases the burden on the centralized data repository thus rendering it inefficient.
Healthcare organizations are expeditiously adapting analytical solutions to maintain systems that are efficient, accurate, maintainable, scalable, energy efficient and safe. Use of edge computing is becoming inevitable
Smart healthcare systems uses wearable, implanted and embedded devices to monitor the health of the patient, wearables being the most widely used. The positioning of smart device is dependent upon the user preference, physician and the device’s healthcare application.