The Inevitability of Cloud Adoption Means A Challenge for Business Infrastructure Maintenance

The existing on-site infrastructure is gradually turning into a relic of the past, akin to the once-ubiquitous floppy disk. David Kaláb from Adastra outlines strategies for navigating the shift towards cloud computing.

Choosing and configuring the right cloud tools has become such a complex discipline that it often requires assembling a team of specialists. These experts select truly efficient and beneficial solutions from a range of available options.

Adastra, a Czech consulting company with extensive expertise in this field, reported revenues of 5.6 billion CZK last year.

“Those who fail to fully leverage cloud capabilities within five years will face difficult times,” predicts Kaláb, foreseeing a bleak future for systems operated on proprietary infrastructure. Innovations in data management are a daily occurrence, and major players like AWS, Google Cloud Platform, and Microsoft Azure are continuously developing their services. The challenge of finding the optimal solution becomes even more complex when considering the addition of private cloud options or on-premise infrastructure operations on one’s own hardware.

In an interview with, David Kaláb shares his team’s approach to cloud solutions.

Is the cloud an inevitable future?

I tell all our clients that those not in the cloud in five years will face harsh realities. They won’t be able to innovate sufficiently, keep up with competitors, respond quickly to market changes, or find people to maintain complex server infrastructures.

Why is it difficult to find people?

There are several reasons. The range of services available in the cloud will continue to grow, while those that can be operated on proprietary infrastructure will diminish. Experts capable of managing older solutions will quickly disappear from the market. The influx of new professionals in this field will slow down as it becomes less promising, and those who remain will be snapped up by IT departments of large corporations, willing to pay whatever it takes to keep their systems running. Consequently, it becomes clear that businesses of all sizes, from medium to the largest, will need to seek solutions in the cloud. Proprietary hardware will become a relic of the past, like the floppy disk, which is now virtually obsolete.

Am I understanding correctly that large corporations won’t be actively transitioning to the cloud?

The entire enterprise sector is gradually shifting towards the cloud. In the banking sector, data analytics are already cloud-based, all telecommunications operators are running many things natively in the cloud, and the shift includes insurance companies, automotive manufacturers, energy sector, and others. Of course, there are higher demands for the security of central systems and user data. However, there are ways to start with relatively low risk – for example, restricting access from the external world, anonymizing data for analytics, or encrypting integrated data from various systems end-to-end. There are indeed many possibilities.

Concerns about cloud security are often mentioned as a weakness.

Yet, if we look through the lens of penetration tests and various certifications, major public cloud providers are currently more secure than any local data center. The key is to find capable people who can correctly set up and maximize the offered security level, whether according to their internal policies or in compliance with regulatory requirements. In professional circles, there’s no longer any doubt about cloud security; the technology has matured, and it’s now up to companies to extract the maximum benefit for their business.

How are the business benefits of the cloud measured?

I would almost say it’s purely a cost issue. The crux of the matter is the return on investment. Companies don’t hire consultants just to help them operationalize the cloud; they hire them to find ways in which the cloud can contribute to business success. Besides return on investment, the cloud can shorten the time from development to the launch of a new service, the so-called time-to-market. In many areas, cloud integration directly impacts customer experience, for example, in terms of interaction speed.

Can you give an example of a popular feature that the cloud can efficiently handle?

Take, for instance, the deployment of an AI-powered chatbot. I can operate it on my own infrastructure, but I need to allocate sufficient computing power for maximum load, i.e., when the most users are accessing it. For a global company, I must ensure deployment and operation in several locations worldwide to avoid long minutes of generating and transferring responses. Deploying and operating such a service typically takes several weeks to months. Alternatively, I can simply launch the chatbot in the cloud, configure it as needed, and take advantage of all the cloud benefits – pay according to consumed performance, use geographically close data centers for the fastest response, and also standard cloud advantages like data backup or rapid scaling of computing power when needed. The time required to deploy such a service can be reduced to a matter of weeks, and if I already operate similar services, even days.

I assume that transferring internal corporate systems to the cloud is not a simple action. What are the biggest complications?

The average larger company tries something with every technology – here, they get tempted by an interesting incentive and unintentionally fall into vendor lock-in; there, they try a promising service that dies within a year, and the resulting gap is expensively filled again. These issues accumulate over time, and towards the end, the company tries to hire a consulting firm for each technology. This is not only expensive, but it also doesn’t solve the problem of an opportunistic approach. What’s missing is an overarching concept.

What do you mean by the incentive that leads to vendor lock-in?

We often encounter clients who couldn’t navigate the wide range of offerings. Cloud service providers offered them huge business incentives to commit to new clients – credits worth hundreds of thousands of dollars, a year of free operation, discounts on services from implementation partners, free migration, and more. Many companies thus signed up for something they wouldn’t normally buy. But once they transitioned to the cloud, any change in provider was extremely costly, so they continued to pay increased fees after the initial discount period ended.

So, in the end, did the cloud end up costing more than if they had stayed on their own servers?

Yes, but often this was also due to inefficient use of cloud services, which is another typical mistake. It’s not enough to just transfer a system or application to the cloud; the entire development process should take into account the completely different concept of the cloud world. Thus, what barely functioned on proprietary hardware could, in the cloud, end up requiring up to a hundred times the computing power of servers around the world. And when the bill for consumed services arrived, everyone was surprised at how expensive the cloud was.

But if I understand correctly, it doesn’t have to be expensive.

Definitely not. The cloud can be operated in a way that doesn’t cost absurd amounts, but you need someone in the company who understands the cloud and can set the right basic principles. Ideally, a company should be able to leverage the cloud’s strength as an operational expense. Today I order and consume, tomorrow I pay. I don’t pay for anything I don’t need or will need in three years. Of course, there are large companies that still try to get costs into depreciation because they report EBITDA (earnings before interest, taxes, depreciation, and amortization) to shareholders and management, but even this can be very well managed in the cloud.

How challenging is it to design a new service operated in the cloud?

Cloud is a demanding discipline. It’s not enough to go through two or three courses on cloud architecture to cover 90 percent of what companies commonly need, as was the case a few years ago. At Adastra, we deploy experts to the project, from architects to specialists in services of individual providers, to the phase where, if necessary, we call in several dozen consultants. Subsequently, the team narrows down, and more technical roles take over to launch the entire solution.

Your recent project for Eurowag won the Data Solution of the Year award at the Microsoft Awards. How does the cloud help in truck transportation?

Eurowag is a beautiful example of a company that has managed to completely turn its thinking around. It’s a progressive fintech company with the strategic goal of becoming the data number one connecting fuel sellers, truck drivers, and the offer of related services, such as its own navigation. This wouldn’t be possible without the proper use of cloud technologies, sensible migration, and the use of new data tools. We helped Eurowag transfer its data platform to the cloud environment. For nine critical systems and over a thousand interconnected tables, we created a solution that can handle even Eurowag’s ambitious goal of becoming a fully data-driven company.

In the cloud, there’s often talk of combining various specialized services. What’s your favorite choice?

Each provider tends to be strong in something different. For example, we often build a data platform on Amazon Web Services, over which we run analytics in Microsoft Power BI for data visualization. This is, for instance, a combination that purely Microsoft or Amazon alone can’t offer, as these two worlds only intersect marginally. But even here, it’s true that services evolve, and in a year, everything might be different.

Multicloud is often mentioned as a defense against vendor lock-in. Is this something you help clients build?

For smaller companies, one cloud provider usually suffices, and multicloud isn’t advantageous due to their size. But for larger companies, it makes sense. If a client doesn’t want to be a cash cow for a cloud provider, they can create their own market where they force cloud providers into competition with each other.

How do they force this?

If their application is indifferent to the environment in which it runs, they can operate services with the provider offering the best price for a given deployment. Or they can build their solution on services operated by multiple public cloud service providers. However, multicloud isn’t just about savings; it can also be a way to minimize risks.

Adastra also operates in North America. Do you perceive any differences between the Czech and American environments in their approach to the cloud?

American companies are much more prepared to innovate. When a new service is released, such as Microsoft’s planned Fabric solution for data analysis as a service (SaaS), Czech companies play with it, ponder its benefits, but no one has yet taken decisive action. On the other hand, I know of several of our customers in the USA, more than the fingers on one hand, who are ready to immediately switch to the new service to ensure they don’t miss out. We try to bring this mindset from North America to the Czech Republic.


The article was originally published on on January 4, 2024.