In addition to software and hardware, proper functioning of the Data Center means a reliable power supply and cooling system infrastructure, and adequate physical security.
According to the Gartner report, many Data Centers are not compatible with modern servers, memory and network infrastructure, and they should be regarded as a living organism that evolves with IT equipment. Even in the standard Data Center there is a problem of high electricity consumption which is 150 times higher in comparison to identical square footage of the office space in modern buildings. In addition, the density of equipment and heat dissipation per m2 becomes bigger and inflicts serious challenges to the cooling system in the Data Center.Hardware that is most commonly used in the cloud infrastructure is greater and significantly more exploited than the infrastructure in regular Data Centers. Cloud Data Center servers consume far more electricity than regular servers, and therefore the power distribution units (PDU), distribution cabinets and uninterruptible power supply (UPS) are exposed to a greater effort. Virtualization leads to a general increase in server load and in the heat emitted by the equipment, which is additionally a demanding issue for the cooling system. The flexibility of transferring the workload within the Data Center and from the Data Center to another can lead to a sudden overload of the power supply and cooling system, and even lead to the failure of the key systems. The Data Center reliability is defined by many standards, among which the most relevant is Tier standard defined by the Uptime Institute. It is the only standard which allows individual certification: project documents (Design), built Data Center (Facility), and management system with the characteristics of the location (Operational Sustainability). Tier Standard is divided into ‘Tier Standard – Topology’ and ‘Tier Standard – Operational Sustainability’. Tier Standard Topology recognizes 4 levels of reliability, which are marked by Roman numerals, with a larger number denoting a better quality of infrastructure:
Tier I – Basic Infrastructure
Tier II – Redundant capacity components
Tier III – Concurrently maintainable
Tier IV – Fault tolerance
Tier levels vary according to the infrastructure configuration, the possibility of infrastructure maintenance and response to failure (fault), as shown in the table below:
Tier Standard – Operational sustainability defines the most important risks that may jeopardize the Data Center, as well as the priority and the most important steps and strategies for their elimination, keeping the level of reliability. This standard complements the previous standard (‘Topology’) and includes:
- Management and administration;
- Facility characteristics of the Data Center;
- The Data Center location.
They recognize 3 categories of management and location quality:
- Gold – top-quality administration and management, with a minimal location risk. Reliability (according to topology – Tier level) within the standards or even exceeds the standards. Recertification is done every 5 years.
- Silver – excellent administration and management, with a minimal location risk. There is a possibility for improvement so that the reliability (according to topology – Tier level) can be achieved. Recertification every 3 years.
- Bronze – minimal administration and management, with a certain location risk. There is a considerable room for improvement. Recertification is necessary every year.
It is not always recommended to plan the highest level of reliability (Tier IV), but it is necessary to match the requirements of a business demand with Tier levels, and maintain business continuity. The ‘cloud’ integration will certainly influence the Data Centers in the future, which will become more modular and more flexible, and adapted to the demands of the processes that are being carried out.
Author: Miloš Stojković, Solutions Product Manager