An essential task of the data center administrator is the planning of the IT infrastructure which depends on the actual and future requirements. For the data center architecture, two aspects are important. One is the level of availability and the other is redundancy. How to build an optimal data center infrastructure and optimize data center efficiency? This article will provide some information for reference.
The following steps can be followed to build an optical data center infrastructure.
Generally, the racks in a server room or in a data center will be used according to the internationally standardized 19-inch format. The front panels of the racks are multiples of one height unit (1 U), equivalent to 1.75 inches. These enclosures are mainly used to stack the built-in hardware in a mechanically secure way and accommodate the cables needed for guiding the power and data in and out. Moreover, they should allow sufficient air exchange for cooling and protect access to the front and back and only allow access to authorized personnel.
The power supply chain is a crucial element in the data center architecture. Depending on the level of availability and resilience required, power is supplied via one or more independent feeds. The low voltage distribution board also distributes power to the various systems and functions of the IT infrastructure, e.g. the different primary supply chains, the UPS (uninterruptible power supply) system and the backup diesel generator.
Much of the electric power fed into the data center is ultimately converted to heat by the IT hardware. This heat must be dissipated from the data center, which is accomplished by generating and distributing sufficient cooling. Cooling can be produced in various ways. Chillers, such as refrigerators, generate cooling by means of electric power, which makes them major energy guzzlers. However, in many cases, free cooling with outside air can drive down the temperature of warm water sufficiently to meet a data center’s cooling needs. Chillers and free-air cooling systems can complement each other very effectively, assuming appropriate control systems are in place. So special attention should be paid to the appropriate cooling technology in a data center, because a major share of operation costs is caused by cooling. Depending on the heat load and the number of IT racks, an adequate cooling technology has to be selected for the data center.
A star topology is recommended to implement a structured cabling solution. With a star topology, maximum flexibility in the network is achieved. The cabling infrastructure should be implemented to allow moves, adds and changes (MACs) without disturbing the cabling itself. MACs include network reconfiguration, growing and changing user applications and protocols.
Generally, modern logical architecture can be broken into four layers—core, spine, leaf and storage. The critical job for many designers is to translate the logical topology onto a structured cabling infrastructure. To translate the logical network into a structured cabling infrastructure, the data center will be segmented based on the logical topology. Each zone should use a middle of row (MoR) interconnect solution for the cabling, and within each zone, the equipment distribution area (EDA) will utilize a top of rack (ToR) interconnect. The EDAs will serve the network, server and storage equipment in each cabinet and the zone distribution area (ZDA) will serve the EDAs. The ZDAs on the other hand will guide the cabling back to the main cross-connection cabinets in the MDA (main distribution area).
Maximum efficiency can be achieved by considering the following key elements: performance, time, space, and sustainability.
To maximize network performance, make informed decisions about the rack, cabinet and cable management solutions that can accommodate higher weight thresholds, and have adjustable rails and wider vertical managers, along with integrated cable and airflow management options for better protection and airflow. The physical solution should support both copper and fiber media.
As data centers are growing in size and complexity, they should be able to adapt quickly and easily to support changing business requirements. Selecting infrastructure solutions that can optimize time, would result in faster deployments, reduced cost, and allow for easier moves, adds and changes. The infrastructure should support heavier equipment, increasing port and cable densities, and cable media changes. A modular solution, which combines the advantages of standardization, provides the foundation for a flexible and scalable building infrastructure.
As space is a premium in the data center, the infrastructure system should be built for optimization. Adopting the rack as the basic building block for data center density is suggested as a best practice. With port densities continuing to increase, making efficient space utilization is even more important. One of the best methods is to use high density connectivity options to optimize space while supporting large port densities.
Sustainability can refer to the ability to last or continue for a long time. Sustainable designs offer choice and flexibility in space design, reducing installation time, material waste on site, etc. They should ensure optimal energy efficiency and performance. This translates into data center solutions that should use a comprehensive approach, active and passive cooling, power distribution, airflow control and racks or cabinets with cable management, to ensure optimal energy efficiency and performance.
Data centers deliver critical services for the business. Building an optimal infrastructure and trying to maximize efficiency are really vital. The suggested steps and elements for better data center performance are only for your reference. Maybe more advisable suggestions can be offered as the technology develops.