Cloud computing or edge computing: what is the best approach?


Data management can be done through two different approaches: cloud computing and edge computing. Which to choose?

2022-04-20T13:45:06+00:00May 19th, 2022|Author: Niccolò Cacciotti |Category: |
Home/Artificial Intelligence/Cloud computing or edge computing: what is the best approach?

Every day we deal with data, from personal information to statistical surveys. So data management is a vital activity, because thanks to their processing it is possible to knowIts acquisition and organization is not directly entrusted to the human being, but to the IT systems. Two are the most well-known and opposed approaches to data management. We are of course talking about cloud computing and edge computing.

Defining the two computing approaches  


Cloud computing represents the first data management approach, while edge computing is a more innovative model that is becoming very popular. They have well-defined features that make them unique. Let’s analyze them one by one.  

Cloud computing refers to the use of multiple services such as software development platforms, storage, servers through Internet connectivity. It takes place without direct management by the user.

In general, cloud computing allows users to store and access data through remote databases and retrieve them on request. Since servers can be located anywhere in the world, the performance of services is not always excellent. Long distances and other hardware and software problems can slow the flow of data, so issues related to latency, bandwidth and in general of the user experience can occur.   

Services are public or private. As for public services, they are provided online for a fee, while private ones are hosted on a network for specific clients. Moreover, this approach typically uses a “pay-as-you-go” model. Thus, users pay only for the cloud services they use. 

In the case of Edge computing, there are two separate versions of this explanation. Let’s discover them.

On one side, through edge computing it is the data processing that happens at the network edge, therefore moving the computation closer to the source. On the other hand, it is possible to move compute power physically on the device where data is generated, for example a smart sensor. It is usually addressed as on board.

This last definition best represents the essence of edge computing. In addition, given that the volume of data has increased, more and more of the processing is occurring at the edge. What is the result?

Wherever and whenever users want to have data available, they can obtain it in real time. This is because information is not managed on the cloud filtered through distant data centres, but on devices. This means that it eliminates delays and saves bandwidth.

Cloud and edge computing: between pros and cons


A question arises: What is the best approach to use? 

We need to understand that cloud and edge computing are different technologies, thus, there is no universal answer. Surely each one is the right choice depending on its peculiarities and business needs.

Therefore, now we just need to find out the strengths and weaknesses of each approach. Let’s start with cloud computing. 

Cloud computing

 Cloud computing

Companies have adopted it so widely because of the benefits it brings. Here they are below:  

  • Scalability and reliability

It allows enterprises to start with a small cloud implementation and expand rapidly.
Moreover, data backup is easier because data can be replicated across several redundant sites on the cloud provider’s network.

  • Mobile Accessibility

It is supported to a greater degree by this computing paradigm.

  • Maintenance

The cloud computing providers conduct maintenance activities.

  • Save expenditure

Companies can significantly reduce their operational costs when expanding their computing IT capabilities.

Businesses that choose cloud computing need to be careful about the possible problems they might face. For instance, its use may cause:

  • Loss of control

Enterprises hand over their data and applications. So they will be totally dependent on the cloud provider.  

  • Problems related to the cloud provider 

In the event of a loss of Internet connectivity from your provider, your business may experience downtime.
Also, if your cloud provider closes or changes its business domain, your company will stop working. 

  • Security threat

Hackers could breach websites such as those of cloud service providers. Without control over data security, you could suffer losses.

What about edge computing?

Edge computing

Edge computing

This is the approach that is growing strongly. It offers the possibility to overcome the limitations of cloud computing and thus achieve better and unprecedented results. Some advantages of Edge computing are presented below:

  • Guaranteed local computation

Internet access is not necessary for edge computing, which can compute and store data locally. Processing takes place in inaccessible locations, making it impossible to stop services for any connectivity disruption. 

  • Speed

Accumulated data does not have to cover huge distances or move at all, as data processing occurs at network edge or on-board. Network  productivity is accelerated by minimizing latency

  • Performance improvement

When the bandwidth of a network is too low, there is a tendency to reduce the size of the data fed into Machine Learning (ML) applications. When at the edge, data feedback loops can be used to improve the accuracy of Artificial Intelligence models. In addition, multiple algorithms can be run simultaneously.

  • Reducing costs

Edge computing provides organizations with more bandwidth and storage at a lower cost than cloud computing.

  • Security 

Processing data within the local network reduces the possibility of cybersecurity attacks in the cloud and thus increases security.

Also in this case, there are a few weaknesses that distinguish the approach. We list the main ones:

  • Cost and Storage 

There is a cost on the local side and also one associated with edge computing when the old IT network infrastructure is replaced or upgraded.

  • Lost Data 

Many edge computing devices discard irrelevant data after collection, as it should. If it is relevant, that data is lost and analytics in the cloud will be complicated.

It seems clear that the disadvantages of edge computing affect to a lesser extent its performance. This approach is chosen for its high efficiency.  Will it therefore be the most chosen approach in the coming years?

In the future: between edge computing and new technologies


Edge Computing is expected to explode in the next few years. Accordingly, more and more enterprise data would be generated and processed “at the edge”.

At the moment, however, there are still many challenges to revolutionizing the entire technology market. Any examples? The integration of millions of devices and software that can monitor and update a potentially infinite number of edge platforms around the world, not to mention security issues.

The role of 5G will be crucial. It is and will be one of the main drivers for edge computing. It allows the increase of interconnected data sources and processing units, making the currently well distributed “cloud computing” obsolete. This is due to the number of connections and data processing requested at the same time in a centralized paradigm. Thanks to this technology, it will be possible to compute an uncanny number of tasks at the edge of the network, thus close to the data source. The outcomes are astonishing for users, particularly in terms of bandwidth and latency.

The rise of applications will need advanced tools able to orchestrate, scale and autonomously manage them. Here comes the second pillar: containers.

Containerization, indeed, will play a major role. This paradigm is a virtualization method where software files (including code, dependencies and configurations) are bundled into a package and executed on a host by a runtime engine. It will provide a way to deploy a large amount of applications across the network, even if placed in remote and distant locations, contributing to scalability, portability and reliability of the solutions. Furthermore containers don’t require a linear increase of costs over scaling to manage working environments.

In the coming years we will see many changes, but mostly evolutions.
Which technologies will you embrace?

© Copyright 2012 – 2022 | All Rights Reserved


Author: Niccolò Cacciotti, Head of AI Department