Why even consider total cost of IT ownership?
IT infrastructure is still an enigma for many companies, yet its selection is a basic economic decision.
It might seem that the purchase of IT equipment – including servers or a cloud account – should fall under the purview of the procurrement department. So, all that would be needed is a stamp of approval from management. But the reality is different. Such a purchase should be considered a long-term expense, i.e. a well-thought-out step.
It’s not just about buying equipment, like computers or phones. A company’s IT infrastructure is an investment that can really cause trouble to a company if its selection is not based on the right principles.
But be careful! These principles are not a universal guide for everyone. They are more like guidelines that companies can use to make their specific selection focused on their specific needs successful.
So let’s look at these principles from the perspective of possible selection scenarios (or dilemmas) and at the same time imagine an illustrative company that gives these principles a clear outline.
On-premise vs. the cloud
This topic is accompanied by endless debate. We are still in a state where one camp is vilifying one option and the other camp is vilifying the other. Neither is correct. The question must always be framed like this: “Is this the best option for our company?” If we leave out the bolded words, the answer loses its meaning.
On-premise solutions are not bad inherently, cloud is not bad inherently. These approaches become bad choices when they do not align with the needs of the company and its direction.
What principles do apply here?
1) We need to find out what options our company has with regard to financing.
- Are CAPEX or OPEX more affordable for us?
- What mix can we afford?
- Is the choice sustainable in the long term?
- Do we have a plan B if this choice changes the conditions for our company over time?
2) Are we aware of all the costs associated with the choice? With an on-premise solution, we need not only capital for the initial investment and finances for the operation of individual elements (e.g.: energy, cooling, space or security), but also finances for management, repairs, incidents and a team that oversees functionality. With the cloud, we are talking not only about the costs of computing resources, but also, for example, licenses, data transfer fees or technicians who are in charge of the entire environment (including security experts).
3) Which option will make it easier for us to ensure business continuity and prepare a disaster recovery plan? Which infrastructure will support us in resilience to outages and errors so that we can minimize the negative impact on our company’s ability to take care of customers?
4) Given the current solution, should we change our approach because we are not utilizing the servers at a healthy rate? How stable is our load? Do our needs include the need to scale quickly and dramatically?
5) Is the solution in line with the level of security and compliance that our internal regulations and the law dictate?
6) Do we even have the internal people to operate such a solution? Can we possibly hire them within the required time horizon and our budget?
Yes, all these questions and issues are the same “boring” stuff over and over again. It would be much more interesting if the choice of infrastructure was guided by some technological fashion cycle. But that is not the case. It is a standard economic decision based on the standard economic principles.
Consolidation and virtualization as low-cost on-premise optimization
Another possible situation is the optimization of the on-premise solution that the company wants to stay with. How do you do that?
First, you need to find out what the current situation is. If physical servers are running at 40% of their capacity tops for a long time, then this is not the most efficient way of operating. Especially if the company operates multiple physical servers with such a load.
Such inefficiency can be further emphasized by one circumstance. If you are purchasing new hardware, but its utilization is still in this percentage range of load, you need to change your approach.
There are several solutions and they do not necessarily involve moving to an external environment. The ideal direction is to choose virtualization or containerization of operations. These procedures help companies make better use of the physical computing resources of a given server. There are several such approaches and they come in different forms, so again it depends on good analysis to reveal the right path.
Cost optimization in the cloud
But what if you have a cloud? Then it depends on how you use the cloud. The more provider-reliant your service is, the fewer options you have for optimization – except for switching to another solution.
So if you use SaaS, your hands are pretty much tied. But you knew this risk when you chose it, so there are probably other reasons to use such a solution.
With PaaS and IaaS, the options are better because you get to the layers that you can work with. But this requires experts in the given environment.
If we are talking about virtual servers, then with some providers it is possible to optimize them by choosing different types of these servers. However, it may not be completely without problems, because some advantageous offers are linked to commitments, others to the instability of the environment.
If a company needs multiple virtual servers for operation, it is worth considering switching to IaaS. This model requires more active involvement of the customer in management and operation, but thanks to this it also gives more freedom in setting up and securing the infrastructure.
The cloud does not aim to be cheap and can never be cheap without work. It is the cheapest it can be, when it is actively and continuously optimized. Various FinOps procedures help with this. However, if such advice leads you to manually turning anything on and off, try to think about whether there is a more suitable solution that will not burden you like this (you can read about this in the case study of DIKA servis).
How to find the right solution for your business?
As one movie says: “I can better show you than I can tell you.” So let’s follow this advice and apply the principles of choosing an IT infrastructure with the most favorable TCO for a given company to METALFAB CZ, a.s. This illustrative company will guide us through the process that ends with the most effective selection of IT infrastructure.
—
METALFAB CZ a.s. with 180 employees is faced with a decision whether to renew on-premise servers or move to the cloud. The company works with proprietary production data that must not leave its own infrastructure.
Starting point
- 9 physical servers, 6 years old
- Utilization: 20–35%
- Annual operation (energy + service): 300,000 CZK
- HW renewal would cost 2.8 million CZK
- Some systems (ERP, design documentation, production parameters) must remain on-premise for security reasons
TCO analysis of different options (5 years)
A) Full on-premise in its current form
- TCO: 5.3 million CZK
- Security is met, but the solution is inefficient, because the company pays for the performance of 9 servers, when it actually only needs 2 or 3.
B) Full on-premise in an optimized form
- TCO: 3.4 million CZK
- Security is fully ensured by maintaining on-premise operations.
- Consolidation of physical servers and their optimization using new technologies.
- External management of layers above the hardware, so the company does not have to purchase its own internal team.
C) Full cloud
- TCO: 6.3 million CZK
- Security is not met, because proprietary data would have to go to the cloud, which would not meet compliance.
D) Hybrid + consolidation (3 on-premise servers + cloud)
- TCO: 3.5 million CZK
- Security met, because sensitive systems remain on-premise.
- Consolidation from 9 to 3 optimized servers, which means the company will save significantly, but does not need experts in a new type of technology.
- Cloud will be used only for non-proprietary IoT and analytical tasks.
What the analysis revealed
Restoring the original solution as it was is a completely ineffective option for the company. Although it provides all the necessary security and compliance, it does not make financial sense.
The way forward could be to consolidate the on-premise infrastructure and use some technologies for more efficient use of physical computing resources. Security and compliance are fully maintained. However, the problem here is the dependence on an external team that will manage the infrastructure set up in this way.
The company can completely eliminate a pure cloud environment because it does not guarantee compliance.
The hybrid option offers a suitable solution that combines maximum savings and compliance with security and compliance requirements.
A clear choice?
So which solution will METALFAB choose? It depends. After all, it should not be purely a financial decision, but mainly priorities that the company has should come into play.
It may happen that one of the priorities is to maintain the way of operation. Then the company will pay extra, but it will fulfill its internal obligation. It may happen that it wants to modernize, so it will choose a hybrid solution. Or the company’s goal may be to gradually switch to a multicloud solution, which will then transform the architecture. Then it makes sense to start transferring some activities to the cloud.
As you can see, economic decisions always have two lines – qualitative and quantitative. Therefore, there is no universal answer to the question – how does one reduce the cost of ownership of IT infrastructure?
