The “cloud” now looks back on a decade-long history and even though it has become part of our everyday lives, it's often still not entirely clear what we mean by it. In this article, we clarify all the fundamental questions, list the best cloud services, and discuss the role and potential of Artificial Intelligence in this location-independent technology.

What is a cloud and what is it good for?

The term “cloud” in digital technology basically just means using remote access and a specifically designed infrastructure during a workflow. For example, when we're not keeping certain data on our own devices, but on servers, which are located elsewhere: this is cloud-based data storage. Of course, any software service can also be performed on these remote devices, so the “cloud” can not only store data but also process it.

The fact that we can do something at a certain distance, even on the other side of the planet, can be useful in itself, of course. But cloud applications are also advantageous for the fact that they don't use our own infrastructure, and we don't have to use our own storage capacity. This way, we don't have to worry about an unexpected power outage, and our data is being stored in multiple physical locations at once, so it remains safe even if a major disaster would occur.

We have already summarized the basics of the “Internet of Things” technologies in this post. This is related to the fact that these IoT devices are almost always connected to some kind of cloud application, they transmit data to it, for example, but they can also be processed there and new instructions can be returned from the cloud (take for example a system that switches on the heating when a resident sets off from work).

okos megoldások a házban, irodában

Our living space is adapting to our habits more and more quickly

Cloud-based systems

The application of cloud technology is already very common even at the user level. Lexunit also uses such tools on a daily basis, like Gmail for emails, Google Drive for remote-access documents, Slack for collaborating between different locations and Wordpress for managing websites. Mobile phone manufacturers offer cloud services for all phones, for data backup for example, which is why it’s so easy to transfer everything from one phone to another, especially if their manufacturer is the same.

While in developers' work, Virtual Machines can be of great help. Google and big tech companies, for example, offer services where we can “borrow” some of the computing capacity of their servers and use it as a virtual machine, both with Windows or Linux operating systems. Lexunit regularly uses Amazon's service, AWS (Amazon Web Services) to create virtual computers, databases, and multiple backups for client projects, and then write the final versions to a physical drive. Such services can be used to model not only a single computer, but entire complex systems, and they operate at very high data throughput rates.

Industrial cloud-based solutions

There are also many examples of the use of cloud systems in Lexunit's industrial client base. In industrial systems, various simulations play a major role for instance. In one of our projects, analyzing the thermodynamic and electromechanical data of turbines, we came to conclusions that could increase the efficiency of the entire system. Here, several different models of physics (thermodynamic, mechanic, electric) run in parallel, and affect each other.

Such well-parallelizable calculations require video cards in high capacities, which can be done via interconnecting multiple cards (clusters), and such hardware is only available from specialized companies. We use such clusters with remote access, meaning that we perform these simulations in the cloud.

Cloud and AI

Today, it's even possible to apply Artificial Intelligence capacities in the cloud. For example, one of the features of AWS mentioned above is SageMaker, which was invented for just that.

With SageMaker, any external partner can build, teach, and implement Machine Learning models. Developing your own ML algorithms is a complex process, making it very time-consuming and expensive. Here, "complex" also means that several different tools, workflows, softwares are needed for its assembly, which is very labor-intensive and gives plenty of opportunities to make mistakes, so a long refinement process is needed to assemble a model that is truly usable. Services such as SageMaker offer these tools and processes in packages. Tasks won't become magically easy per se, as compatibility and coordinated operation must be ensured with SageMaker, but it's often a simpler solution than developing your own Machine Learning system starting from the ground up.

Every major cloud provider from Google to Microsoft Azure already has a similar AI toolbox. These can be used to overcome problems that would inevitably arise if a company were to start developing independently from scratch. Even the biggest brands are taking advantage of this opportunity, from industry through healthcare to recruiting. It can be useful wherever you have to make precise decisions based on large amounts of data.

Let's see some specific examples from the case studies of AWS:

ADP is a multinational HR management company. Using AWS, they became able to analyze various labor market processes almost in real time, allowing them to run models that forecast the impact of a certain rate of salary increase or predict the scale of employee attraction of their client companies based on patterns.

It is well known that the biggest American sports leagues might easily have the most thorough statistics. The Baseball League is using AWS to enhance this by digging up newer and newer, interesting metrics.

“Microgrids” are small power plants connected to the network, such as wind farms or solar panels, that can provide energy at micro-regional levels. The company Advanced Microgrid Solutions builds such systems. One of the keys to developing them is harmonizing the fluctuating energy production with the similarly fluctuating energy consumption. With the help of AWS, they were able to build a Deep Learning model that could improve predictions, increasing the efficiency of the whole system.

természetes energiaforrások felhasználása

AI can play a major role in energy management

Reuters is one of the largest news agencies in the world. They wanted to develop an app for their clients, where they can ask questions and get relevant answers based on their massive news database. The implementation of this classic Natural Language Processing task required the capacities of SageMaker.

And even Tinder… yes, even this app uses Machine Learning tools to decide who to show as a possible match, and SageMaker helps out with that too!

Data flows upwards

In general, cloud technology is increasingly moving towards outsourced hardware rental, meaning that soon we can access machines with extremely strong computing capacities with simple user consoles - if we have broadband internet and, of course, if we pay for rental fees. This way, the tools around us can become smarter, more sophisticated and more efficient, as their “brains” can potentially be somewhere else, under the supervision of experts, supported with continuous optimization.

Are you interested in starting a similar project but would like to ask some questions first?
Make the most of our free consultation service and contact us today. Click here.