How to choose a data centre?

Raul Kallaste
Head of Business Development of the MCF data centre

Kristo Keiv
Sales Manager of Datafox


The choice of a data centre is often based on traditional criteria of choice, such as duplicated power supply from special substations, manned security, and fire extinguishing systems that are friendly to IT equipment. However, in fact, a number of other indicators must also be taken into account. A data centre is not just a space for hosting servers – it is a place to which services offered to end customers move, and their effective performance is the key to successful operation of any company.

• Location: what kinds of buildings are located in a close vicinity of the data centre? If, for example, there is a plant producing hazardous chemicals or a fuel storage facility located nearby, for a company to place its IT solutions into such data centre would most certainly be considered a risky idea.

• Accessibility: how conveniently and quickly the company’s IT team can access the data centre, whether the building will be accessible 24/7 in case of a crisis situation, does it have a sufficient number of parking places, etc.

• Network and service provider neutrality: In data centres of big ICT companies their customers often have to use data communication and IT services of the same service provider. If cooperation with their IT teams for some reasons does not go smoothly, the customer effectively does not have any options to choose from.

• Duplicated data communication connections: having two different network connections is certainly not enough – a data centre should have direct connections with at least two domestic Internet service providers, and with at least two international data communication providers. In order to ensure an operating network connection in case of failures, it is important that all connections enter the data centre through different physical cables and network units.

• Working conditions at the location: how comfortable and safe it is to deliver expensive servers to the location – or maybe the stairs are so narrow that equipment can be carried inside only manually? Also, there is a big difference between a situation where in order to set up a server the IT specialist must sit on his haunches, or where he can use an internal network of the building and comfortably work at his office desk. Comfort of work is also influenced by the internal temperature – whether in the server room it is kept at 23 degrees, or the operator prefers to optimise cooling costs, and work must be performed in a 43-degree heat.

• Safety for the environment: an environmentally friendly approach becomes an increasingly important consideration when buying different ICR services. Therefore, it would be a good idea to assess whether the data centre uses green electric power produced from renewable sources, whether residual heat generated by servers is used for heating, and what other measures for reduction of carbon footprint and preservation of the environment are used.

• Understandable price formation: it is important that all costs are immediately clear for the customer, and that later he does not discover in the invoice any mystical additional services such as obligatory purchase of a car parking place, additional charge for night access, or installation charge, even if the customer has to carry his server cabinet in his own.

• Full service support: whether it is necessary to bring servers and other devices into the data centre, to set up a virtualisation environment, operating systems and databases, to ensure cyber security and backups, and to provide monitoring and system administration. Without those components a cabinet with devices standing in the data centre would be no more than just a very expensive fridge. This is why before making a decision it is a good idea to inquire, whether the administrator of the data centre is able to provide those services on its own or with the help of a reliable partner.

• Comprehensive monitoring solution: data centre service providers often offer a monitoring solution that only shows whether the server responds to ping requests or not. This is not enough, as the risk of hardware failures is significantly smaller than the risk of failures of IT solutions provided to the end customer. Therefore, the monitoring system must cover the entire IT solution, showing whether all functions of the service are working properly.

• Hardware and software purchase model: if a customer has to purchase all IT infrastructure on its own, it means the need to make quite big investments. Therefore, the customer may want to prefer those data centres that offer service-based IT solutions (ITaaS, IT as a Service), where servers, other devices and software licenses are paid for on the basis of a monthly fee.

Importance of every criterion of choice depends on specific needs, however, the safest option would always be a data centre that offers a comprehensive service. In such case, if IT needs of the customer change, it will be easier to change the content of the services without having to go through relocation of the entire equipment elsewhere.

 Back to articles