Serverless computing is a model based on an architecture that does not require server management to build and run applications. Consistent with the new breed of cloud services that facilitate the use of innovative technologies, serverless computing offers significant cost, efficiency and productivity gains, says Jonathan Hiben, Software Architect and Technical Lead at Ainos.

A wide range of services 

"Cloud computing consists of renting remote servers rather than buying them as was done previously", Jonathan Hiben recalls. "The advantages of cloud computing include the fact that the user company is no longer responsible for maintaining its infrastructure: it is the service provider who is in charge of providing perfectly functional computers", he adds.

Moving from a capital expenditure model to the operational leasing of IT equipment can be disorientating at first, but proves beneficial in the long run by allowing to fine tune the use of resources. "This model has been broken down into specific sub-sets, from the simplest - renting a 'bare' computer - to the most abstract, depending on the supplier's assumption of an increasing share of capacities and functions, with the corollary that the company can fully focus on its core business", Jonathan Hiben comments.

Cloud computing allows organizations to have applications, computing power or storage resources as shared, dematerialized, contracted (from a performance, security level, and cost prospective), scalable and self-selected services. "The public cloud giants posses immense capacities in terms of resources deployment and economies of scale", Jonathan Hiben observes. "This allows them to provide servers and abstraction solutions at very attractive prices".

"On the left end of the spectrum", he explains, "we have Infrastructure as a Service - or IaaS, a service where the provider is only responsible for making secure machines available. Then there is Platform as a Service - PaaS - which gives the user the freedom to modify certain parameters such as computing power, memory, network or code deployment, but where the notion of physical server is starting to fade. At the other end of the spectrum, we have Software as a Service - SaaS - which offers the use of a complete software package on a rental basis, as is the case with Microsoft 365, for example. In this context, the customer has virtually no responsibilities other than ensuring that user accounts are properly managed".

Taking full advantage of the cloud's elasticity

Serverless computing is a cloud-native model that allows developers to build and run applications without having to manage servers. "Contrary to what its name might suggest, the serverless model does rely on servers, but they are almost invisible to customers", says Jonathan Hiben. In fact, the cloud provider takes care of the routine work, supplying the server infrastructure, keeping it running smoothly and scaling it up. Developers then only have to load their code into containers to deploy the applications.

"Even if the serverless architecture can present different gradients of user intervention, the interest of this model lies in its capacity to automatically allocate and calibrate the resources necessary for the proper execution of workloads, whatever the demand may be", Ainos' Technical Lead explains. "With this in mind", he says, "we can envisage Function as a Service, or FaaS, i.e. a cloud service that enables a function to be triggered following a given event. At this level of abstraction, any notion of computing power or memory disappears completely". Once serverless applications are deployed, they respond to demand and scale automatically when needed.

In general, serverless offerings from public cloud providers are charged on the basis of this event-driven execution model. Therefore, when a function is idle, it costs nothing. "In the case of serverless, and even more so of Function as a Service, the price to be paid for a given service is really correlated to its actual usage", Jonathan Hiben confirms. "The more a service is used, the higher the price. Conversely, when this service is little used, the amount the customer has to pay decreases sharply".

"The benefit of the serverless model is twofold", he says, "in that you can deal instantly with an increase in demand without maintaining an oversized infrastructure and benefit from a sharp reduction in costs when demand is at its lowest. For example, at one of our customers who bases part of its activities on a serverless architecture, the resources are only used during actual working hours, whereas during the night or lunch break, consumption - and therefore the price to pay - is at its lowest. This model makes it possible to take full advantage of the elasticity of the cloud".

Serverless computing applies not only to code, but also to databases or certain API management tools, for example. "At Ainos", says Jonathan Hiben, "by choosing a serverless architecture for the development of a CAPTCHA test prototype, we were able to start the project with a modest investment while benefiting from the possibility of scaling up the solution when success was achieved. The serverless approach allows for greatly reduced time-to-market", he underlines.

Interview by Michael Renotte

A specialist in Microsoft technologies, Jonathan Hiben is a full-stack developer, capable of working on the applications front-end, databases and back-end. He holds a dozen Microsoft certifications: Azure Engineer Associate, Solution Architect Expert, DevOps Engineer Expert, but also AI Engineer Associate and Data Engineer Associate.

Through his experiences, Jonathan has acquired skills in software and solution architecture (solution design, microservices integration). He is also involved in the pre-sales phase of projects and participates in the development of offers.

As a team leader, he supervises a team of 7 people and ensures their career development and knowledge maintenance. At the same time, he is Technical Lead for one of Ainos' clients.