Data Center Terminology: 12 terms you should know


You may have heard administrators in the data center mention some strange terms - what do they all mean? This overview will help answer that question.

ASHRAE stands for American Society of Heating, Refrigerating and Air-Conditioning Engineers. It is a commercial group, but the relevant departments are Technical Committee 1.5 (Computer Application) and Technical Committee 9.9 (Important Facility, Data Center, Technology Space and Electrical Equipment). death).

Blade servers are modularized servers, designed to save space and serve fast maintenance. In a typical rack, hundreds of blade servers can fit into space for dozens of standard servers.

Close-coupled cooling - Link cooling mode is installed right next to a rack. This is useful for a rack with components where a standard cooling system from your data center is not enough.

Clusters are networked server clusters. Modern server clusters can be remotely distributed, resulting in a virtualized or software-defined cluster. Regardless of the setup type, a server cluster is useful for applications that require multiple servers or caches running in parallel.

Containers are separating areas in your operating system reserved for certain applications. Containers are often used in virtual operating systems. Applications tend to be collapsed in the form of microservice. Docker, RKT and Mesos are common containers. Kubernetes is an open source program for container management.

Disaster Recovery - DR is a disaster recovery system, which involves (begins with) off-site data backup (outside the main data center). The idea is to get your data center back up and running as quickly as possible, even if it's a temporary or minimal configuration, while your main datacenter is repaired or rebuilt after a carpet. art. Many service providers offer DR, and like all other services, you'll want to test it to make sure it is serviceable.

Hybrid Cloud is a system in which one or more public cloud services are combined with resources in your own data center, into an almost entirely managed environment. That sounds great in theory, but the management can be extremely complicated. Ask questions if your vendor or system integrator advertises "hybrid cloud solutions" with words like "easy" or "simple".

Hyper-converged - Hyper-converged is a term under discussion: Ask 10 people and you will probably get 20 answers. Technology analyst Zeus Kerravala explained: “There was a hyper-converged infrastructure market that lacked the software aspect when the technology came along. Hyper-converged platforms must be turnkey products that include all hardware and software, a miniature data center in a box.

PDU - Power Distribution Unit is a power distribution unit in data center. Sometimes the term is used to refer to the electrical system of your entire data center, but most often people refer to the electrical system inside the rack. PDUs can be either vertical bars on a rack or they can be focused on a common rack unit. Sometimes these components are combined with uninterruptible power supplies.

PUE - Power Usage Effectiveness is the effectiveness of the use of electricity. It is an industry standard (ISO / IEC 30134-2: 2016), which measures the ratio of energy used to energy supplied. Perfect PUE would be 1. Most modern data centers have a PUE of about 1.2-1.4. Whenever this parameter is above 1.5, you need to find a solution to improve the power problem.

Software-defined is the art of separating the network part from your hardware. You can use virtualization software to create, change, scale or delete networks defined by the software as you like. Networks can be distributed or combined. You can even allow applications to do this for you, based on your needs and policies.

Tiers, as defined by the Uptime Institute, certify four data center levels. From the Uptime Institute website: Tier-1 is the most basic classification and Tier-4 is the highest fault tolerance. A data center provider called Switch defined their own Tier-5 in 2017.