"Data center for beginners" or criteria for evaluating the project of an effective data center

Many companies create their own data centers, and an even larger number of companies (and private individuals) rent either third-party data centers or equipment in such data centers. Reliable IT infrastructure now plays one of the most important roles in the success of most companies.

A large number of external and internal factors are forcing companies to develop and expand such infrastructure. Some factors, for example, the expansion of the company's business, lead to the need to expand the data center, other factors - the need to increase the reliability of the security of their data.
')
A good data center project with its subsequent implementation is the key to successful work of tens and hundreds of thousands of modern companies, both domestic and foreign. What are the criteria for evaluating an effective data center?

There are not so many of them, and all of them, for the most part, are clearly dependent on each other.

Reliability

In this case, reliability means the ability of all systems to perform their functions as planned, within a certain time. The degree of reliability of the data center can be calculated using such indicators as the mean time to safe operation (MTBF, Mean Time Between Failures) and the average time to recover performance (MTTR, Mean Time To Repair).

It goes without saying that to create a truly efficient data center, you need the appropriate equipment, including reliable power sources, cooling system components, servers, data storage systems, etc., and so on.

Availability = system uptime / total running time (work + idle)

As already mentioned in Habré, the availability of a data center is classified as follows (one of the most popular classification systems):

• Tier 1: 99.671% Availability Guaranteed
• Tier 2: 99.741% Availability Guaranteed
• Tier 3: 99.982% Availability Guaranteed
• Tier 4: 99.995% Availability Guaranteed

The design of the data center should take into account the availability of the most important elements of the system, including the servers, storage systems, applications, cooling systems and so on mentioned above.

Criticality

Another important parameter that must be taken into account when planning data centers is the criticality of applications and services supplied to the end user. If the service is very critical, and its simple amount will cost a decent amount of the company, the data center and the entire infrastructure must be built accordingly, guaranteeing a high degree of system availability. If the service / application is not very critical, and the company has little to lose from more or less lengthy downtime, the availability of the data center may be lower than in the previous case.

Criticality is desirable to assess on a 5-point scale, where 5 is the maximum criticality, and 1 is the minimum. Financial and banking applications / services should be evaluated, for example, as 5. Accordingly, all the elements of the data center should be in this case not only reliable, but as reliable as possible.

Capacity

The volume of the data center, its capacity is an important parameter, since the amount of equipment that can be placed in a certain volume will depend on it. This is an important parameter when planning sufficiently productive cooling and energy supply systems.

Development plans

Another important factor that is sometimes not taken into account when designing a data center. For example, if a company plans to develop, it is worth planning a larger amount of data center space than is currently needed.

Otherwise, the leadership over time can get a headache in the form of the problem of the impossibility of expanding the current capacity, despite the vital need for such an expansion.

Scalability

This factor is “tied” to the previous two. All data center systems: information storage systems, network infrastructure, power supply, cooling, everything should be planned based on the need for scalability.

Scalability is divided into vertical and horizontal. Horizontal scalability - increasing system capacity by adding nodes to the cluster. vertical scalability - increasing the computing power of the system by replacing the entire system or individual modules of the system with more productive ones.

Efficiency

With the emergence of all the new "green" technologies, such as energy-efficient equipment, the efficiency of the data center can be gradually increased. However, it is desirable to fix this indicator during the planning process.

Equipment density

For each type of equipment, this indicator has its own value. Usually the density of the equipment remains more or less unchanged. This parameter is tied to availability, efficiency and scalability.

Serviceability; usability

It all depends on the initial project, as far as all possible problems have been taken into account. Ideally, any problem that arises should not cause problems with access to equipment by the repair team, for example. In a good data center, almost all the problems that arise should be solved “on the go,” without stopping the entire system.

Design, hardware, service systems, the entire architecture of the data center should be planned on the basis of usability, ease of maintenance.

Controllability

And finally, another important factor that shows how easy the data center and individual systems are to manage. This factor affects the ease of conducting operational tasks, monitoring and analysis of work.

What else?

In general, several other factors can be identified, however, those mentioned above are the most important.

Of course, with one size all projects can not be created, because any data center is unique. But the criteria for evaluating the effectiveness and reliability of the data center project should be uniform.

By the way, someone can call the ideal, in his opinion, data center - his own, or someone else's company? It would be interesting, and not only to us :)