Guide: How to calculate the benefits of migration to the "cloud"

Disclaimer : A more detailed analysis and detailed analysis of the formulas is presented in our thematic book .

The topic of the economic rationale for the transition to the clouds remains relevant despite the fact that in specialized literature and in the Internet media it is usually written about it rarely and sparingly. On the one hand, the model of technical and economic calculations itself is quite simple and boils down to an ordinary calculation of one-time and operating costs.
')
On the other hand, collecting accurate input data for correct calculation is not an easy task at all. And if you use incomplete input data, then the result of such calculations will be far from the real state of affairs.

Therefore, cloud providers try to avoid rough calculations (“by eye”) in specific projects, and for promotional purposes they use examples that are rather flawed in terms of data completeness, illustrating the “effectiveness” of using cloud solutions. In this material we will examine the working models for the calculation of economic efficiency and give examples of such calculations.

As a basic cloud service, we consider virtual infrastructure lease services in the IaaS (Infrastructure as a Service) model. Read more about what constitutes such a model, can be found in the book " IaaS for business by brick ".

It is important to understand: technical and economic calculations are indicative. They should be carried out if your task is to justify the expediency of upgrading the IT infrastructure at a specific object (and not “on the fingers”, but on real numbers) and choose the most effective (including from an economic point of view) implementation of the IT infrastructure alternatives.

Selection scenarios

The need to make a feasibility study of moving to the cloud usually arises in two cases:

• there is a need to upgrade the existing infrastructure;
• New resources are required for a project, planned expansion, or it is planned to create an IT infrastructure from scratch.

Enlarged IT infrastructure can be divided into three parts: the site, equipment and virtualization environment. Each of these parts can be outsourced along with lower level tasks.



Levels of IT landscape transfer to outsourcing

In this case, the main options for comparison in terms of cloud availability (or outsourcing) can be:

• purchase of equipment and placing it in its own server;
• purchasing equipment and placing it in the data center;
• equipment rental in the data center;
• rental of virtual capacity in the data center (the actual cloud in the IaaS model).

Within the framework of these options, various architectural solutions are possible, providing, respectively, different availability of IT infrastructure at different costs. Therefore, it is incorrect to compare the economic efficiency of different “as-is” infrastructure options. To do this, it is necessary to clarify the IT infrastructure readiness ratio required for the business and bring each of the options to the required value of this parameter.

We understand the terms: what is the availability factor

Availability is an important indicator that allows you to compare two or more options for the implementation of IT infrastructure. Comparison of two options, for example, the cost of ownership, will be incorrect with different availability. In other words, the availability of IT infrastructure is the probability that our infrastructure will be in working condition at any time (except for the windows of scheduled maintenance).

When determining the required availability factor, the following data are taken into account:

• volumes of direct financial losses in idle business applications;
• indirect losses caused by the inoperability of business applications (decrease in the level of trust, transition of customers to competitors or abandonment of services, etc.).

Based on these indicators, we can talk about the maximum allowable downtime of the entire IT infrastructure and individual services. The approximate calculation of the availability factor is as follows [here you can refer to the material of V. Alekseev, “ Approximate calculation of the degree of readiness of the system ”]:



where A is the system availability ratio;
t _p - the total time the object is in working condition;
t _in - the total time to restore the object.

To find the necessary parameters, you need to do the following steps:

1. To make an architectural scheme of the system.
2. Convert it to logical.
3. Break into modules with serial / parallel connection of components.
4. To perform the calculation of readiness for the modules.
5. To perform the calculation of readiness for the system as a whole.

---

The approach in which we make a start from the required availability factor is convenient, first of all, for projects on creation of new IT infrastructures. In this case, we first determine the required availability ratio, and then consider the options for implementation that provide this ratio, and compare them on the cost of ownership.

If our task is not to create the infrastructure from scratch, but to modernize the existing one, then two ways are possible:

1. We are satisfied with the current availability ratio, we are looking for a solution that will allow us to ensure a given level of availability of IT services with minimal costs. In this case, we need to achieve a positive internal economic effect, and the external one may be zero. Options with the same availability and identical functionality can be compared in terms of cost of ownership.

2. We consider several solutions with higher availability or better functionality; in this case, the internal economic effect will be insignificant, perhaps even negative. In this case, the main savings should be achieved at the expense of external economic effect - for example, by reducing the losses from the downtime of IT services. The effectiveness of several options with different availability factors can be compared in terms of the annual economic effect (SER).

We understand the terms: the economic effect - internal, external, annual

The most convenient indicator of the economic efficiency of IT infrastructure used in practice is the annual economic effect (SER), measured in rubles per year.



where e is the annual profit (savings) achieved when modernizing the IT infrastructure at an economic facility [rubles / year];
K - capital (one-time) costs for the creation and implementation of IT infrastructure components [rubles];
E - rate of return (standard profitability) [1 / year].

Annual profit (E _year ) in can grow both by saving various cost items (costs), and by increasing revenues from the introduction of cloud technologies.

Capital (one-time) costs (K) for the creation and implementation of IT infrastructure are the investments necessary for the creation and commissioning of computing power. A detailed list of possible capital expenditures with the formulas for their calculation is given in our thematic book .

The rate of return (regulatory profitability) E is the rate of return on capital and the rate of entrepreneurial income. Its value is directly related to the standard payback period for capital investments T _ok :



The value of regulatory profitability is selected based on the situation in a particular industry and in the economy as a whole. In general, the value of E should not be less than the deposit rate of a reliable bank (otherwise, instead of investing in the modernization of the IT infrastructure, it is more profitable to simply deposit this money in a bank).

In the IT field, the pace of the emergence of new technologies and software and hardware upgrades is such that their lifespan is usually considered to be about 5 years (after which they, in theory, are subject to write-off or significant modernization). For this reason, the value of E in the calculations for the IT infrastructure, as a rule, should not be less than 0.2–0.25 (otherwise, the investment in it will not be repaid during the actual operation).

When calculating the indicator of the annual economic effect (SER), it is customary to subdivide it into two components: internal (direct) and external (indirect) economic effect. The internal (direct) economic effect is achieved due to immediate savings in the implementation of the functions of providing the object with the required IT services.

In general, we can talk about the following internal effects:

• Reduced downtime of business processes due to failures;
• reducing the risk of losing business-critical data;
• increase productivity by increasing the level of mobility of employees.

The external (indirect) economic effect is achieved mainly by increasing the efficiency of the company's main activity as a result of reducing losses from the downtime of information systems and from the loss of information important for business.

Features of calculating the cost of ownership of IT infrastructure

If we need to compare different IT implementation options, similar in terms of availability rate (for example, IaaS and creating the appropriate infrastructure on our own), we need to take into account a number of features.

Hardware redundancy and computing power

To ensure a given infrastructure availability ratio, hardware redundancy of servers may be required. If a company implements the IT infrastructure itself, then hardware redundancy translates into unused computing power reserved for the failure of one of the servers. Equipment downtime in this case may occur not only because of the need to ensure the required availability factor - planning to increase the load in the future, the company usually buys equipment “with a reserve”.

As part of working with an IaaS provider, the reservation is implemented at the hardware level by the provider itself and is already included in the cost of the service, and the increase in computing power is not discrete (“hopping” —as the case with the in-house infrastructure), but almost continuously. Accordingly, there is no talk about unused (and, thus, unprofitable) powers.

Investment risks

For an investment project in the IT infrastructure, as for the IT department as a whole, negative risks are primarily characterized, which consist in the probability of losing resources with unjustified investment. In this case, if we consider the capital expenditures in the IT infrastructure as an investment, then when comparing with IaaS, where there are no significant capital expenditures, it may be necessary to adjust the calculation of the cost of ownership.

Such an adjustment can be implemented by calculating the notional cost of insuring each of the risks relevant to a particular project: for its approximate calculation, it is necessary to multiply the amount of possible losses by the probability of a negative event.

Carrying out such calculations is quite a painstaking task: in order to calculate the real benefits from the implementation of a particular format of IT infrastructure, you need to take into account many factors that economists are used to working with, and not IT specialists. Nevertheless, such a calculation can be much more convincing than many stories about clouds and success stories of companies using the services of IaaS-providers: in most cases, the benefit from migration is expressed not just “in words”, but also in real numbers.

PS: A more detailed analysis and detailed analysis of the formulas is presented in our thematic book .