As a huge number of companies are relying on Data Centers part of their business and digital information, the Data Center infrastructure provider must assure a certain level of security and guarantee in terms of redundancy.
Redundancy in data centers is defined as a system design where a component is duplicated so that in the event of any power outage, equipment failure or any unexpected breakage, the IT equipment will not be affected in an alarming way. A good example could be having power redundancy in case of an outage.
The maximum tolerable period of disruption (MTPD) is decreasing because most companies are having less tolerance for their business and operations. ITIC (Information Technology Industry Council) released an annual survey explaining that 98% of companies said that 1h of downtime could cost over 100,000$.
The redundancy approach in Data Centers
The most evident concern for Data Centers is to avoid downtime, and that is usually the driver for data center redundancy. The approaches are, mainly, data center power redundancy, cooling redundancy and multiple telecom entrances
We can talk about Data center power redundancy when two or more utility feeds, generators, UPS systems and outlets connect to each rack.
Although it is rare, a device can fail, leaving you with an outage of service while the flow of power is interrupted. With redundant power in our Data Centers, you minimize this risk of this happening by connecting to an independent power system within the data center.
The truth is that this redundancy approach in Data Centers minimizes downtime associated with both failure and the need for maintenance.
N+1 vs 2N+1 Concept
- N means the duplicated unit (CRAH unit, diesel generator, UPS…), so a N type design means no redundancy. In case of any failure, some main equipment and services would be affected.
- N+1 means that one additional item is added as back up in case of failure. A common design implies 1 extra unit every 4 needed.
- N+2 is a completely redundant system. Every item is not connected with each other and is completely independent.
The TIER classification is a standardized methodology according to the official regulation ANSI/TIA-942:
- TIER I: Dedicated Infrastructure. 99.671% uptime.
- TIER II: Redundant Infrastructure. 99.741% uptime.
- TIER III: Fully Fault-Tolerant. 99.982% uptime.
- TIER IV: Fully Fault-Tolerant. 99.995% uptime.
Basic data center with no redundant components. This is a suitable design for business that tolerate planned shutdown periods.
Redundant data centers (N+1). These ones include all of the TIER I capabilities and add extra redundancy for critical equipment such as power equipment (UPS, generators…) and cooling systems (HVAC, pumps and chillers).
Redundant data centers (N+1). These ones include all of the TIER I and TIER II capabilities and add extra redundancy. The equipment is connected to more than one power and cooling networks, with only one active. This is suitable for business than can’t tolerate large shutdowns due to the extremely importance of data stored.
Annual availability of 99.9995%. Maximum annual shutdown of 52.56 min. The systems are connected to multiple power and cooling networks with multiple redundant equipment, following the 2(N+1) criteria.