You can't avoid the cloud. We advise companies to take a sober approach and start with proof-of-concepts, says David Kaláb of Adastra

• Will all companies end up in the cloud?
• How to sell the idea of a cloud solution to a CEO?
• And if you decide to migrate, what strategy should you choose?
• Vendor lock-in - is it something to worry?
• What are the most significant differences between private versus public cloud? 

Ivana Karhanová: Is it possible to avoid the cloud, especially if you are a large corporation with hundreds of thousands or millions of customers? Or should you be prepared to overhaul your entire infrastructure from top to bottom? And how do you know that you really shouldn't invest a penny in your existing solution? I'll be talking about it with David Kaláb, who heads up the Government, Utility & Insurance division at Adastra. Hi.

David Kaláb: Hi.

Ivana Karhanová: Is it possible to avoid the cloud?

David Kaláb: Yes. If you say it like that. But the question is different. Do we want to avoid it, especially in the future?

Ivana Karhanová: Do we want to, or do not?

David Kaláb: I think that if we look at business parameters, we cannot avoid the cloud in the future.

Ivana Karhanová: Does that mean that all companies will end up in the cloud, whether they want to or not?

David Kaláb: When we look at smaller companies today, they are already there. Whether it's different ERP systems, CRM systems, etc., very few people are running own systems today. Most of the smaller companies are using cloud platforms, and the question is when the big ones will use them.

Ivana Karhanová: Your division deals with government, utilities, or the insurance segment. These companies have hundreds of thousands and millions of customers, and I assume they may be flirting with the cloud or trying something out. But many systems are still running on-premise.

David Kaláb: That's right. And these big companies don't have it easy. The environment is much more heterogeneous; they run many systems that are often customized - tailored to individual customers. They also have large volumes of data, complex processes, and so on. For them, of course, it's going to take longer, but even there, it's a matter of, say, a few years. Leaving aside regulatory or security constraints, even these companies will gradually use cloud services more and more.

Ivana Karhanová: What is the current view on using the cloud? I assume that they will always deal with some cost-benefit analysis.

David Kaláb: Exactly. That's the biggest snag we see with companies. However, it's not appropriate to look at it as a business case that compares the current cost of cloud services with some hardware and licenses. Instead, it needs to be looked at through the company's goals - how quickly we want to respond to changes in the market or our service portfolio, how quickly we want to serve customers, and so on.

For example, as part of our strategy, we set out that we want to be able to launch a new service or product within three months. By the time we can implement it and put it into production, we already know that we definitely won't be able to do it in three months. So we need some prebuilt solution just to use, integrate and scale these things. So that we don't have to test everything the way we try it today on-premise. If we build it that way, we'll find that cloud services may be the only answer.

Ivana Karhanová: How do you sell this idea to the CEO?

David Kaláb: As I said, if I have a standard analysis done comparing the cost of a data platform, this business case doesn't work. You have to realize that my existing hardware or licenses will become obsolete at some point, and I will no longer be able to keep up with the business because I can no longer service its requirements. And if I only make decisions when that happens, it will be too late.

Ivana Karhanová: So what strategy do you and your team advise to choose to make it OK for everyone?

David Kaláb: Even for larger companies, we think it's good to start playing with the cloud if it's not already happening. Start implementing individual proof-of-concept (POC) solutions for the different services I want or run to get a sense of how it behaves, scales, how DevOps works, etc. In short, to be prepared internally and externally. If I can't do it internally and don't have the people, then hire externally. In short, to be ahead of the business so that I'm ready because there won't be time to test anything the moment the need comes.

Ivana Karhanová: The operation of cloud services is quite different; for example, if the on-premise server is running 24/7 or is "unplugged" for an hour, nothing happens. However, when the cloud is running 24/7, the bucks pile up...

David Kaláb: That's a governance thing - how I want to build the cloud and how I want to run it. It's also about changing the mindset. In the cloud, the situation is different than with your infrastructure. I can add or subtract performance on an ongoing basis. If I'm running internet banking, for example, I can say I need to increase performance at eight in the morning, whereas I don't need any big performance at midnight, and I can reduce it and save money. Compared to existing infrastructure, it is more challenging in this regard. Still, on the other hand, I'm able to adapt over time in the cloud, which is very difficult in on-premise infrastructure.

Ivana Karhanová: If I understand correctly, the flexibility that I have in the cloud, on the other hand, is offset by the fact that I have to be careful where what is consumed so that I don't have deployed machines; left running that nobody needs anymore?

David Kaláb: That's right. When I look at the companies already running more extensive solutions today, there's been an awakening everywhere. In the beginning, the goal of cloud providers is to make you as a customer consume as much as possible. So hence, they paint the services pink for you to try. Then you run it, and you calculate that it will cost you, say, $30,000, and eventually, it comes out to $60,000. So you start to optimize and somehow figure out the other side.

Ivana Karhanová: When companies decide to move to the cloud, what strategy should they choose?

David Kaláb: I already touched on that. It would be best if you started building some internal pilots. At the same time, you need to bring in people from companies that are already doing these things because these things change a lot over time. It's a matter of trying it out, planning your strategy from the bottom up based on how different POCs work out so that you don't get surprised one day. If I know in the future that I have an ERP somewhere, a CRM somewhere, and I have a piece of on-premise, so that I don't have, for example, data flowing up and down ten times and I'm surprised at how much it's costing me in total. That's something to think about as part of the multi-cloud strategy that you've chosen.

Ivana Karhanová: The takeaway is that you need to be incremental. For example, we will deploy everything new in the cloud and gradually leave it on-premise.

David Kaláb: That's the ideal approach versus the greenfield approach, where you say, OK, now I'm going to containerize all the development and applications and move the data entirely to the cloud. That's not entirely realistic. If you do it incrementally, you're probably going to have some surprises. Still, it's about mindset and trying to stop looking for arguments why it can't be done and instead consider the business benefits in the future.

Ivana Karhanová: We've already mentioned the cost of the cloud, which companies used to on-premise may not expect at all. Is it a good idea to sit down with someone first and discuss the whole architecture and how it works, and that maybe it needs to be built a little differently?

David Kaláb: Absolutely. You need a skilled business architect at the beginning to think about what the core requirements are, the core processes, and so on. Then, it would be best to have someone who will find a counterbalance in running the whole thing - a technical architect who knows how to integrate that and set up security. At the same time, you need a DevOps engineer to help set it up, handle development, releases, and so on.

Ivana Karhanová: What are the most notable differences between private and public cloud?

David Kaláb: Private cloud is often used by companies to eliminate the cost or responsibility of taking care of their hardware and licenses. However, it's a cloud by halves. Usually, when I go to a private cloud, I'll take a server at home, move it to the private cloud, and pretend that it's a piece of my infrastructure. The change is about moving from infrastructure services to platform and service services. For me, it's not so much private cloud versus public cloud, but more whether I'm still looking at it as infrastructure or whether I'm able to consume services as a platform or a native service.

Ivana Karhanová: I see the most significant benefit of public clouds: vast teams of developers behind them, access to the latest and most efficient solutions from people I probably never have a chance to hire and get into my business.

David Kaláb: That's right. However, when I look at it as a bank, it's a slightly different perspective than a commercial company. As a bank, I want to make sure that it's mine and that nobody will touch it and see it. I want to be in control of it, and at the same time, I have obligations to the regulatory authorities. On the other hand, if I'm a commercial firm that needs to launch a new product, I can benefit from having experts, services, information, or documents available to help me accelerate my development. So there are different motivations for doing it one way or another. However, it is wrong if I do it the other way, i.e., if I build a custom solution in a private cloud, even though I don't have to. By doing so, I make the whole thing more complicated just for the false sense of being in control.

Ivana Karhanová: What about vendor lock-in with the cloud?

David Kalab: That's an interesting topic. I think it will be mentioned more and more in the future. In the beginning, companies may not perceive it as much, but the ambitions of individual cloud providers are clear. They want to get companies quickly, and they can provide various considerable incentives to do so. So once they get you and get, for example, 50 percent of your core applications into their cloud, it's more or less clear that you won't migrate elsewhere.

Ivana Karhanová: Is vendor lock-in the same as on-premise solutions?

David Kaláb: It can be, but there is a counterpart. The moment I submit to that, I can do things relatively quickly. If, on the other hand, I tell myself that I want to be immune to vendor lock-in, I'll build it so that the services are as similar as possible across clouds, and if I get annoyed with one vendor, I'll go to another. It has to be said that, at least for the time being, this approach will make it more expensive.

Ivana Karhanová: So what do you advise companies?

David Kaláb: A sober approach. Unfortunately, today you probably can't avoid choosing a path you're going to take and hoping that it doesn't backfire. In the future, maybe when services are more commoditized across cloud providers, you'll have more opportunities to migrate elsewhere.

Ivana Karhanová: Says David Kalab, who runs the Government, Utility & Insurance division at Adastra. Thanks for coming into the studio, and see you sometime.