Telecommunications Industry Association recognized by ANSI (American National Standards Institute) has categorized datacenters based on infrastructure availability as Tier-1, Tier-2, Tier-3 and Tier-4.
Looking at the datacenter availability perspective , Tier-1 datacenters availability is 99.671% whereas component capacity would be non redundant. Tier-2 data centers have the redundant component capacity with infrastructure availability 99.741%. Tier-3 data centers meet Tier-1, Tier-2 requirements along with multiple paths including the dual power path and infrastructure availability of 99.982% and Tier-4 would cover all the above features along with dual cooling equipments such that heating ventilating , air-conditioning (HVAC) cooling system and fault tolerant infrastructure along with infrastructure availability of almost 100%.
The datacenter power usage effectiveness is the unit to determine the energy efficiency. The commonly used formula to measure the power usage effectiveness is Total Facility Power divided by IT equipment power. Power is the most essential component while designing the datacenter huge cost is incurred.
Datacenters have huge hunger for power as it uses a lot of power. There are two main categories of devices are biggest consumer of datacenter power. One is the actual component such that Servers and systems , Storages and Network components . The another one is cooling system which is used to cool the actual components. But, actual components are energy efficient. Cooling cost is major one while designing datacenter because it’s going to cost you more .
Data center infrastructure management
Data center infrastructure management is most essential hence entire Information Technology is dependent on it. We have to put the right software and tools to ensure the datacenter power and efficiency. DCIM is used to identify the risk and fix to improve or ensure the availability of the IT infrastructure.
Design concerns based on Service provide to customer :-
Based on service provide to customer we can consider to design the datacenter for example if you are going to design data center to the individual customer we can have freedom to place the data center components. If we are sharing the service to multiple customers than we have to take cares while doing datacenter design. For example if you are providing service to multiple customer through single device it’s very difficult to get the maintenance window , if the component used to provide service to single customer its fine we can easily get the maintenance window else we are dependent on each and every customer whoever getting the service from the said device. It’s tedious to get the maintenance window multiple customers. Hence, each and every customer would have the different business model they would not be able to provide the maintenance as the same day if others provide the maintenance window.
Considering that, we can design datacenters based on the requirement I mean the devices can be grouped based on customer requirement . While designing datacenter based on customer requirement ( if the requirement is small scale ) is it has some disadvantage such that we have to put many number of small unit of devices to fulfill the requirement also we should consider the future enhancement should be put in place else plan must be frozen. But , looking at the current trend of data usage do you think we can freeze the infrastructure growth . The answer is no that’s why Infrastructure optimization through virtualization, analytics to find out the needs , cloud computing and dynamic infrastructure plays major roll in contemporary datacenters.