Buying or renting: expert opinion

Buying or renting: expert opinion

Michael Blinov

To date, businesses are investing more than ever in the digital infrastructure needed to stay competitive in an increasingly dynamic and rapid environment
the changing economic landscape. Investments in the purchase of server and computing resources, network architecture and cyber-security tools will vary by industry,
budget and goals of the company.

Given the rapid pace of development of the IT environment, one of the important questions that companies need to find an answer to is whether a physical server, a virtual infrastructure are suitable for solving their business problems.
or a combination of the two?

While physical servers are a proven and powerful solution for creating data centers, it’s still a solution of the past. Modern technologies allow to speak about
virtual infrastructure as a service of the present and the future, as the cloud is focused on rapid innovation, and most importantly – a managed approach to solving the absolute majority of problems
virtually any IT service.

Considering the benefits of these solutions, it is difficult to decide on only one option. Today, many organizations choose a hybrid approach and use a local or a virtual one
infrastructure, depending on the type of information being processed, business continuity and disaster recovery plans, and regulatory requirements.

Ultimately, the companies themselves must choose the solution that will maximize the benefits of their business. It takes time, careful analysis to understand this
existing IT infrastructure and understanding of the direction of development of the company. However, in this article, we can try to compare the cost of owning physical equipment with the cost of consuming
cloud services and try to answer the question that is still economically attractive to businesses.

Consider two options for the example of one large agricultural holding company.

Option 1 – IT solution:

1) Hardware and hardware infrastructure on Hewlett Packard equipment:

  • 5 HP 460c blade servers;

  • SRS with a total volume of 17 280 GB;

  • switch – 2 pcs;

  • HP Hardware Performance and VMware Virtualization.

2) The company has a 25% discount on the vendor.

3) The company has its own server to host the equipment.

Option 2 – Cloud service:

VMware Virtualization:

  • vCPU – 325; RAM – 1526 GB, HDD – 12 250 GB.

Let’s start with the final cost of these two models:


The cost of iron in this case is as follows:


Software Cost:


Operating cost:


Costs to continue software technical support:


The comparison is made for a five-year term. The advantage of virtualization in the form of a gradual increase in leased resources is explained by the fact that they are increasing as needed. In this case
it is planned to launch virtual capacities from 40% in the first year and to full capacity by the end of the third year. This approach allows for a more flexible approach to cost planning than
maximum utilization of computing resources is ensured. While the purchase of equipment usually occurs in an amount that covers not only the current (for a month or two in advance), how much
one or two years ahead. Including an increase in the volume of purchases allows you to get some discounts from suppliers.

To evaluate both options, we use the “net discounted cash outflow (MTPL). It is calculated as follows: the amount of net cash outflow (NCO) includes
operating expenses (OPEX), year-end residual value, tax payments. Next, we need to understand what the cost of OCTs is for the entire project over the five years to date. because
future money is always cheaper than today’s money, so we adjust the value of the OCTs to adjust their value, resulting in the OCTs being projected for a total of five years.

In our example, the OCT was calculated for both the In-House and the cloud, and the difference was 38.5 million rubles in favor of the cloud solution. In five years the equipment will remain
It will be fully operational, but in three years it will be fully depreciated and technical support will become more expensive. And the resale of such equipment involves a discount of more than 50% and not always


A few more important nuances need to be eliminated.

1. When designing and purchasing equipment, provision should be made for high availability clusters. The load on the system is usually calculated from a forecast of one and a half or two years and then
recycling is gradually increasing. But to get a volumetric discount from the vendor try to buy immediately. Additional equipment is taken into account with peak consumption.
In the cloud, however, you can flexibly vary the power and backup is already there: virtual machines are able to move when failures of specific hosts. On the other hand, if you expect
geodistributed cluster for critical processes, will need a second cloud. When using the cloud, it is possible to flexibly scale resources over time to increase seasonal load, or
consume on the Pay-as-you-go model. You can create machines, operate them, and then delete them when the need arises. You can use the service under another billing model and receive
a fixed pool of resources with a fixed monthly payment. Convenient for planning future expenses.

2. The organization of the server will require the cost of either the extension of the office or the rental of a rack in the data center (site with the provider).

3. Operation, namely the payroll fund, is rarely considered right. Not included are training with new “iron” and software and additional specialists. These are taxes, contributions to funds, LCA,
cellular communication, costs of workplace (stationery, computer), rent of the area, part of the salary of the staff.

4. Cooling, fire safety do not directly relate to IT, but have a cost element. The price of power is dependent on the consumer, and for an office kilowatt costs on average more than for a factory
or the data center.

5. Not only servers and GDGs, but also network components are depreciated. In a normal server, the term of depreciation is three years, in a switch – five years. After that the equipment must be changed: it grows
number of failures, the need for repair increases. The equipment not only produces its own resource, but also becomes obsolete.

6. In the above calculation, the customer chose to compare with HP servers because he had experience in operating them. Of course, you can find a cheaper Chinese counterpart, but then it will lead to less
functionality of the equipment, difficulties in its installation and operation, less elaborate technical documentation. If you refuse support or take support with a weakened SLA, then
there will be risks of downtime and losses.

7. In the cloud, you can lease Microsoft licenses (or software services) and not buy licenses for years to come, but pay them for actual use on a monthly basis. Or move some
your Microsoft licenses and do not incur additional costs if the provider has a signed Mobility license agreement.

Using virtual and physical servers has its pros and cons, and considering the features of using them, as well as the economics of ownership, will help you understand which option is best
suitable for business.

Cloud: More flexibility, less control

What are the advantages of cloud over iron? Business card for cloud infrastructure – convenience:

  • Someone else maintains a public cloud infrastructure. You do not need to administer hardware in a public cloud. The cloud service provider builds new servers, replaces the old ones
    machines, faulty components and ensures the relevance of security mechanisms.

  • Technical support – 24/7. No need to worry about firmware updates or replacement of faulty hard drives. OS patches are posted automatically with PaaS and SaaS (in fact, you
    no worries about OS at all), removing the load from your IT department.

  • The cloud is flexible. Cloud service can be stretched as much as necessary, responding to an increase in computing power. If you need a site for several months, you can quickly and quickly
    easy to deploy and disconnect as soon as the need for it disappears. Such promptness was named the main reason for the rise of clouds in the 2014 Gartner survey.

  • Cloud scaling. If all or most of the operations are performed locally, scaling usually requires the purchase and deployment of new servers, which is expensive and
    a long process. In the cloud, you can pay as you use it, and it scales instantly (sometimes compared to utilities). If the organization has a seasonal peak
    activity, you can simply add cloud resources or user licenses as load increases, and reduce volume when normalized.

  • Setting up a cloud is easy. Getting started with an IaaS solution is as easy as filling out an online form and clicking Submit, depending on the service. Compare this to the speed of equipment purchase,
    debugging and installation of servers, as well as further support.

The cloud has many advantages, but this does not mean that it will suit every organization. One of the main factors keeping companies from using cloud technologies is
the fact that the data is on someone else’s servers, perhaps in another city or country, and there are legal restrictions in certain areas, such as healthcare, finance and the public sector.

Locally: More control, less scalability

What are the advantages of iron over cloud?

  • Local server performance is more stable. At its core, cloud computing is IT resources delivered over an IP network (for example, via the Internet). Therefore at low speed
    Internet connections may be delayed, which may result in complaints about slow server performance and reduced performance. An internal approach to calculations means that data is stored
    locally (there are no long distances for transmission over the Internet), so access to them is fast and without leaving a private network.

  • You can access your local data offline. The Internet-oriented nature of the public cloud means that ISPs can pose technical problems
    company in a difficult position. This is not a problem for local servers, as you can access files even without an Internet connection and physically go to the server if
    something is wrong.

  • The choice of local equipment depends on the company. You have the freedom to buy and upgrade your equipment.

  • Security is the responsibility of the company. With its own infrastructure, the company knows where its data is sent and where it is stored. No need to send confidential
    information over the Internet, and it will not appear on the hard drives of any server in a random data center – control remains.

  • Local costs can be more manageable. As the company grows, economies of scale can also work in its favor, potentially making local servers cheaper than cloud.

Ultimately, the choice between cloud and internal IT infrastructure depends on the specific needs of the organization. Combining on-premises and cloud-based services can help
get the best of both worlds by finding the right combination of control, cost savings and scalability. In the end, it all depends on the choice and understanding of the needs of the company.


IT-Expert Magazine [№ 01/2020],

Leave a Comment