Cloud

Interview: Regional and open-source-based clouds in Europe

Published on

Clouds of Europe

"Regional open-source clouds are a crucial alternative to Big Tech," says Jurg van Vliet of the Utrecht cloud engineering agency Aknostic. The CEO is one of the driving forces behind the Clouds of Europe project. The goal is to develop regional and open-source-based clouds in Europe. A discussion about the setup and opportunities of regional clouds where Kubernetes plays an essential role.

Aknostic director Jurg van Vliet states that Europe's dependency - and certainly the Netherlands - on cloud infrastructures from a handful of American suppliers is significant. "And that while these suppliers cannot even comply with European legislation regarding privacy and personal data. Is that what we want as Europe? An important question, if only because there is indeed an alternative."

Aknostic was active for several years as a technical consultant in AWS-based cloud environments. Nowadays, they mainly focus on open-source clouds. Van Vliet says, "As cloud engineers, we design, build and manage cloud environments for development teams. With the introduction of AWS in the first decade of this century, a world opened up for software developers. Suddenly the dependency on an IT department was minimal, and we could start unlimited servers via an API and a credit card. This innovation has had an incredible impact on the software industry. During the same period, Google and Microsoft also came up with a similar cloud approach. There are more providers now, but only Alibaba can keep up with these three big players."

Van Vliet acknowledges that it is an interesting business to help companies and governments migrate their IT infrastructure to an environment like AWS. "Until the moment came when we realized that we could build cloud infrastructures for our customers that are at least as good and maybe even better, but where the customer - and thereby the Netherlands and Europe - is not dependent on a single supplier."

This idea was further developed within the Sustainable Digital Infrastructure Alliance (SDIA), a 'not-for-profit organization' that focuses on making digital infrastructures sustainable and inclusive. Within the SDIA, this idea was called a 'regional cloud.' The idea was that data centers from one's region or country are perfectly capable of hosting a complete cloud stack, without necessarily needing a hyperscaler.

An approach that looks much more at local or regional parties using open-source software like Kubernetes. But also asking what such a regional cloud yields for the regional economy and society. How many jobs are created when companies and governments use a regional cloud hosted in their own city or region, instead of, for example, at AWS or Azure? What does this yield in tax revenues for those cities and provinces? Or what about knowledge development - does it stay in the region, or does all that knowledge disappear to other countries? How many jobs are created at the suppliers who provide services or products to these regional cloud projects?

The city of Amsterdam, among others, has been persuaded by the answer to these questions to support the idea of an Amsterdam cloud. Other European cities and regions are also showing interest. The question is only: how do we technically set up such a regional cloud?

Clouds of Europe

Jurg van Vliet, Aknostic (photo: Annelien Nijland)

The term 'regional cloud' has been replaced by 'clouds of Europe,' explains Van Vliet. An English term because this idea is relevant for every city and province in Europe. Central to a European cloud is Kubernetes. This container orchestration platform was originally developed by Google but is now being further developed as an open-source project within the Linux Foundation.

An SDIA document states: "The Kubernetes community is so large and so active that it is becoming the standard. Software suppliers such as Elastic also distribute their software via Kubernetes. So it is now becoming increasingly easy to run the extra services that make Big Tech so interesting yourself."

"With Kubernetes, companies can build their own cloud, which is then offered as a public-cloud solution to other companies and governments. They can host their own applications there, similar to how they do it now at AWS or Azure, but then in a European cloud. As more software suppliers participate (a matter of time), the own cloud can start competing with Big Tech in the eyes of software developers. And the software developer is ultimately 'the boss,' he adds.

Use cases?

"Kubernetes is a very stable basis for a cloud environment" Implementing Kubernetes may be getting easier, but it is indeed a challenge for many IT departments. "That's why we see many technical consultants focusing on this market. I'm not preaching to the choir here, because if we actually want to set up dozens or hundreds of local or regional clouds in Europe, we need at least as many technical advisors. What I hope is that more and more companies will develop a European cloud for the city or region where they are located and offer it as a European public cloud to other companies and institutions in their geographical area."

Kubernetes can be implemented bare metal or on top of OpenStack. "This creates the basis for our own cloud. What else is needed to actually run workloads in the containers that become available depends on the specific requirements and wishes of the customer. When it comes to the cloud infrastructure itself, we can add numerous options for monitoring and other things. Such tools are available as open source. In other cases, it might be a closed-source product, but that's not a problem. The point is that we can build a full-fledged cloud environment based on a standard foundation - Kubernetes - to which the customer's important extras are added. This can be European software, but it doesn't have to be, as long as we host that regional cloud in our own region."

Is Kubernetes as an open-source project mature enough to handle this great diversity of use cases? Van Vliet answers this question with a resounding yes. Kubernetes is very mature, he believes, can handle the load that hyperscalers need, but is also perfectly capable of creating smaller cloud environments. "It is a very stable basis for a cloud environment."

Weak point: environmental impact

However, Kubernetes lacks something, according to Van Vliet, a point that is particularly crucial in Europe. That is insight into the environmental footprint of the applications and the cloud infrastructure built and hosted with Kubernetes. "Almost every developer I meet wants to have insight into the energy consumption and environmental footprint of his or her work. How can I as a developer get insight into how big that environmental burden actually is during the development of my software? And how can I adjust my code so that the environmental footprint becomes smaller? I myself was active for a while with building cloud-based IoT projects for horticulture. Energy use is always very important there because many IoT sensors and devices are powered by batteries. Being very energy efficient is therefore crucial. We can also apply that kind of knowledge to further reduce the energy consumption of a cloud infrastructure."

It doesn't stop at just energy consumption, Van Vliet explains. "As a developer, I want to know and improve the total environmental footprint of my software. So I also want to know more about - for example - the 'embedded carbon' of the materials and components that make up the physical servers and storage devices my application uses."

Finding relevant data Read also E-waste hangs like a shadow over the ICT sector Young companies rake in billions Getting this kind of data is not so easy, Van Vliet says. "Fortunately, a lot of research is being done. For example, the VU in Amsterdam has a research group Software and Sustainability (S2) led by Professor Patricia Lago. Furthermore, the Cedaci research project has provided a lot of knowledge about the materials and components that make up IT equipment. In the Netherlands, we also developed the Server Idle Coefficient (SIC) through the LEAP project, which can be calculated by any party certified by an independent foundation. This allows us to determine how efficiently we use server hardware. But also look at a project like RePlanIT, which develops dashboards to determine the degree of circularity of an IT infrastructure."

"There are many more interesting projects that we at Aknostic are keeping a close eye on," says Van Vliet. "We would like to extend Kubernetes with tools that make all this information available to the developer. So that every software developer gets a better grip on the environmental footprint of an application, virtual server, or container. This will immediately meet a prerequisite for successfully setting up European regional clouds."