The international standard for data center performance was created by the Uptime Institute almost three decades ago. Over the past 30 years, the data center tiers rating system has developed from industry terminology into an international standard for data center uptime. It is important to remember that these standards are objective and should be followed in all applications.
The each tier in the tiering system matches a particular business function. Each one has a standard for maintenance, power, cooling, and fault capabilities. You can utilize the system to estimate the level of security needed and how much potential downtime your infrastructure could experience throughout the year.
Data Center Tier Ratings
For the global standard, Uptime Institute’s methodology, data centers are classified as Tier 1, Tier 2, Tier 3, or Tier 4. Tier 1 is considered the poorest performing classification, while Tier 4 is the highest level a facility can achieve. Since data center technology changes constantly, technology and brands do not matter in the classification. The only factor that determines a facility’s tier is performance.
Tier 1 Data Center
Again, these data centers are the poorest performing amongst the tiers. In a Tier 1 data center, you have one source of power and one source for cooling. If one of these fail, your entire infrastructure could go down. These have a very small capacity and are typically network closets that are the infrastructure for a small business or office. Tier 1 will protect against human error, but rarely unplanned outages.
Tier 1 Data Center Requirements:
- 99.671% availability
- Annual downtime of 28.8 hours
- No redundancy
- One path for power and cooling
Tier 2 Data Center
Tier 2 data centers are similar to Tier 1, except they have at least partial cooling and multiple power redundancies. Fail-safes and equipment in Tier 2 facilities include UPS, chillers, pumps, energy generators, energy storage, heat rejection, and fuel tanks. Unexpected shutdowns and outages can still effect Tier 2 infrastructure.
Tier 2 Data Center Requirements:
- 99.749% availability
- Annual downtime of 22.7 hours
- Partial redundancy
- One path for power
Tier 3 Data Center
Tier 3 data centers are where the largest step up in performance occurs. More comprehensive redundancy capabilities are required to meet this classification. Tier 3 facilities do not require a shut down when proactive maintenance or servicing is needed. This is the first level that gives you multiple days of backup power.
Tier 3 Data Center Requirements:
- 99.982% availability
- Annual downtime of 1.6 hours
- Fully redundant (N+1)
- Power outage protection for 72 hours
Tier 4 Data Center
Tier 4 centers are fault tolerant, meaning that when individual equipment fails or interruptions occur; there is no effect on the infrastructure of your organization. They have several isolated or completely independent systems to maintain power continuity. Every piece of equipment in the facility must have fault-tolerant power design to meet Tier 4 classification. They also require uninterrupted cooling. Tier 4 data centers are used with organizations that need constant uptime, which is typically required with most businesses now.
Tier 4 Data Center Requirements:
- 99.995% availability
- Annual downtime of 0.4 hours
- Fully redundant (2N)
- Power outage protection for 96 hours
You can see how the Uptime Institute’s data center tiers classifications can help an IT professional quickly identify the infrastructure and what service is needed. In summary, Tier 1 facilities will normally have to shut down for maintenance and equipment will fail during outages. Tier 2 data centers have at least partial cooling and have additional fail-safes for power, but still will shutdown during an outage. Tier 3 data centers are able to sustain continuous power during an outage for at least 3 days. In Tier 4, the highest performing data centers, all components are supported by two backups and have outage protection for at least 96 hours.
