Each tier corresponds to different levels of equipment and availability.
• TIER 1 designates data centers with only one lane for the power supply and cooling, without element redundancy. These centers have an availability of 99,671%, corresponding to an average cumulative downtime of 28.8 hours per year.
• TIER 2 brings together data centers that also have a single channel for power supply and cooling, but with redundant elements to achieve availability from 99,749, or 22 hours off.
• TIER 3 has several supply and cooling channels, only one of which is active. Elements are doubled and maintenance can be frequent without stopping the machines. Uptime is 99,982%, or 1.6 hours of downtime per year.
• The TIER 4 is the most demanding: it has several active channels in parallel for the power supplies and cooling. Many infrastructure elements are doubled and "tolerant failures". Uptime is 99,995%, or 25 minutes of downtime per year.