What is Edge Computing?
Edge computing is a paradigm of distributed computing that brings data and computational storage adjacent to the place where it is required to develop save bandwidth and response time.
Edge computing’s origins lie in content delivery networks. These networks were developed in 1990 for serving video and content through edge servers that deployed close to the users. These networks evolved in 2000 for hosting application components and applications as well on the edge server. As a result, the initial economic services of edge computing that hosted various applications like ad insertion engines, real-time data aggregators, shopping carts, and dealer locators.

Significantly, current edge computing develops this approach by virtualization method that enables it to run and deploy a huge application’s range over the edge servers.

The first definition is a kind of computer program. It distributes low latency adjacent to the requests. Karim Arabi broadly described edge computing as every computing external to the cloud proceeding on the network edge and significantly inside the applications in which real-time data processing is needed.

In Karim Arabi’s definition, cloud computing Implements over big data but cloud computing Implements over instant data which is produced by users or sensors.
According to the “Edge report’s State,” edge computing focuses on servers within the adjacent proximity to the end mile network.

The nodes of the edge applied for game streaming are called gamelets that are two or one hope usually away through the client. Edwin and Per Anand say ‘mostly the node of an edge is two or one hopes away through the mobile client for meeting the constraints of the response time for real-time games’ within the context of cloud gaming.


The IoT device’s increase on the network edge is generating a data amount to be computed on data centers and pushing the requirements of the network bandwidth to the limit. Against the network improvements technology, data centers can’t guarantee acceptable response time and transfer rates which can be a complex need for various applications. Moreover, constantly devices on the edge consume data arising from the forcing companies and cloud to create a content delivery network for decentralizing service and data provisioning.

In the same way, the objective of Edge computing is to act the computation aside through the data centers to the network edge, network gateways, mobile phones, or accomplishing smart objects to performs operations and gives services on the cloud behalf. It is possible to facilitate IoT management, storage management, service delivery, and content caching resulting in good transfer rates and response time. At a similar time, sharing the logic within the different nodes of the network defines new challenges and issues.

Security and Privacy

The distributed behavior of this paradigm defines a shift within the security schemes applied in cloud computing. Data may move between distinct distributed nodes linked by the internet. Thus, it requires unique encryption mechanisms separate from the cloud. The nodes of the edge may resource stained devices and limiting the selection in security method terms. Furthermore, a shift through a centralized top-down structure to the decentralized trust infrastructure is needed. Besides, it is possible to move collected data’s ownership by keeping the data on the edge from various service-providers to each end-user.


It must face distinct issues in any distributed network. First, scalability must hold account heterogeneity of various devices, including different energy and performance constraints. It also includes the highly powerful reliability and condition of the connections, as compared to various strong cloud data center’s infrastructure. 

Furthermore, security needs may define further latency within the communication among nodes which can slow down the process of scaling.

Failover management is necessary to maintain the services alive. When an individual node is unreachable and goes down, users must still be capable of accessing the services without any interruptions. Furthermore, edge computing systems should facilitate actions for recovering from any failure and warning the user of the incident. 

All devices should manage the network topology of the whole distributed system to this objective, hence that detection of recovery and error become applicable easily. Other aspects that may affect this factor are the connection method in use that may facilitate distinct data accuracy and reliability levels generated on the edge that can be unreliable because of the specific environmental situations.

Artificial intelligence and sophisticated analytical tools can execute on the system edge because of the analytical resources proximity to various end-users. This placement on the edge supports to enhance operational efficiency and add several benefits to the system.
Edge computing additionally can be used as the intermediate stage among wider internet and client devices outcomes in efficiency savings. For   example, the client devices need intensive processing computationally on video files for performing on outer servers. 

By utilizing servers located over the local edge networks to implement those computations. Various video files require to be sent within the local network. Ignoring transmission on the internet may result in bandwidth savings significantly and thus increases efficiency.

Edge computing delivers analytical computational resources adjacent to the users and helps for speeding up the speed of the communication. Significantly, a well-developed edge platform will outperform a cloud-based system (traditional). A few applications depend on response times enabling edge computing a more possible option compared to cloud computing. 

For example, applications including human perception like facial recognition that commonly takes a human among 370-620ms to performing. Edge computing is capable to mimic a similar perception speed as various humans that are helpful in applications like augmented reality in which the headset must preferably recognize who an individual is at a similar time as a wearer does.

Network Edge

The network edge can be defined as where the local network or device involving a device and interacts with the internet for internet devices. The edge can be a little fuzzy term. Such as a processor or computer of the user in the IoT camera could be granted the network edge, however, the local edge server, ISP, or router of the user are also granted the edge. 

An essential result is that geographically network edge adjacent to any device, unlike cloud servers and origin servers that could be far through the devices they interact with.

Applications of Edge Computing

The services of edge applications can decrease the data volumes that should be moved, the distance and traffic that data should travel. It reduces transmission costs and facilitates lower latency. Computational offloading for many real-time applications, like facial detection algorithms, illustrated considerable developments in response times that proven in various researches. 

Further research indicated that using machines (resource-rich) known as cloudlets close to mobile users. These cloudlets offer services commonly found inside the cloud, facilitated improvements inside the execution time if some actions are offloaded to any edge node. Besides, offloading all the actions may lead to a slowdown because of the transfer times among nodes and devices. Hence, based on the workload the optimal configuration could be described.

Another architecture’s use is cloud gaming. In cloud gaming, a few factors of a game can execute in the cloud but the accomplished video is sent to some lightweight clients executing on devices like VR glasses, mobile phones, etc. Such kind of streaming is called pixel streaming.

There are also other remarkable applications available include home automation systems, smart industry (Industry 4.0), smart cities, autonomous cars, and connected cars. 

Usage of Edge Computing
Most of the companies analyze, manage, and store data over centralized storage, commonly in a private cloud or public cloud environment. Although, traditional cloud computing and infrastructure cannot meet the needs of several real-life applications. Such as, in the case of IoE (Internet of Everything) and IoT (Internet of Things), a highly accessible network along with less latency is needed to process big data amount in real-time that is not possible over classic IT infrastructure. The edge computing benefits are obvious in this case.

Difference between cloud computing and edge computing
The primary difference between edge computing and cloud computing is where the information is being processed. The data is analyzed, processed, and collected in a centralized location in cloud computing. Besides, edge computing is dependent on the distributed computing environment where the data is locally analyzed, processed, and collected. There is no requirement to select between edge computing and cloud computing for many cloud solutions. Edge computing and cloud computing do not compete with one another, they only complement one another and implement together for providing better performance over applications.

Benefits of Edge Computing

  • Speed: All milliseconds in any company are vital for the business. As an outcome of latency and downtime could cost them along with many dollars. The term edge computing includes the ability to improve the speed of a network by decreasing the latency. It decreases the gap travel by data processing adjacent to the information source. The final result is that the latency will not measure in milliseconds but microseconds. Thus, the responsiveness, quality, and speed of the entire service will enhance.
  • Security: The data available on the cloud can be easily hacked. On the other hand, edge computing sends only corresponding data to a cloud. Edge computing sometimes doesn’t need any network connection. Even when a hacker control to infiltrate a cloud, not each type of user data is at risk. But, it can’t ensure that edge computing is risk-free completely. As compared to the cloud, it has fewer risks potentially.
  • Reliability: Edge computing manages reliability very well. It doesn’t rely on the servers and internet connection but it facilitates the uninterruptible service. Various users don’t require to worry regarding slow down internet connections or network failures. Furthermore, it can locally process and store data by applying microdata centers. Because of this reason, it can be ensured that it is a reliable connection for IoT devices. Hence, edge computing is suggested to be applied in remote places where no reliable network connections are available.
  • Cost: Maintaining IoT services can be expensive because for the requirement of more computational power, data storage, and network bandwidth. Applying edge computing for various IoT devices permits users for replacing data centers with the solutions of the device and reducing the bandwidth. Hence, there is a cost reduction in IoT application and device implementation. In inclusion to that, each data is not sent to any cloud. The information will filter and only some relevant ones can be transferred to a cloud decreasing the bandwidth of the network. It can overall decrease the cost of the entire infrastructure.
  • Scalability: The data requires to be sent to any centralized data center in the infrastructure of cloud computing. Most of the time expanding and modifying the datacenter could be expensive. But, the edge could be applied to scale our IoT network without requiring to take care of the requirements of the storage. Furthermore, IoT devices could be expanded here in individual implementation.

Drawbacks of Edge Computing

  • Incomplete Data: Edge computing just analyzes and process information partial sets. Other data are only discarded. Because of this, various organizations may finish losing much valuable information. Hence, the organizations should decide what kind of data they are responsible for loose before applying edge computing.
  • More Storage Capacity: Edge computing hold a higher storage capacity on our device. Hence, the storage devices are more compact. It will not be an issue. But, it is a fact to remember if developing the IoT device.
  • Maintenance: Edge computing can be defined as a distributed system, unlike the centralized cloud architecture. It means there are various combinations of networks with many computing nodes. It needs a higher maintenance cost compared to the centralized infrastructure.
  • Investment Cost: Realizing the edge infrastructure could be complex and costly. It is because of the complexity that requires additional resources and equipment. The IoT device along with edge computing arises with the requirement of local hardware for them for functioning. It can overall cause more efficiency, but a powerful investment is needed.
  • Security: Often, ensuring capable security could be challenging within the edge distributed platform. It is   because information processing takes place on the outside network edge. There are also risks of recognizing cybersecurity breaches and thefts. Additionally, it will enhance the opportunity for many attackers to penetrate the device whenever an IoT device is included here.