The Uptime Institute’s tier system classifies data centers into four levels based on performance, redundancy, and availability, from basic to fully fault-tolerant. A Tier I (basic capacity) facility has single paths for power and cooling, while Tier II adds some redundant components like UPS modules and generators. Tier III (concurrently maintainable) includes multiple, independent distribution paths, allowing for maintenance without taking the system offline. Tier IV (fault-tolerant) is the highest level, with 2N or 2N+1 redundancy to withstand single-failure points in critical systems.
Tier I: Basic capacity with a single path for power and cooling; requires downtime for maintenance.
Tier II: Adds some redundant capacity components (e.g., UPS, generators) to improve reliability over Tier I.
Tier III: “Concurrently maintainable,” meaning maintenance can be performed without interrupting operations due to multiple, independent distribution paths.
Tier IV: “Fault tolerant,” with a fully fault-tolerant infrastructure (2N or 2N+1) that can withstand any single-point failure.
The system focuses on the performance-based goals rather than specific technologies, and each tier builds upon the previous tier
Related:
Bibliography
Reliability Data Collection Dot Std 3006.8: Framework for Establishing Goals for New Data
University of Michigan Cloud Services
Stanford University Cloud Transformation
University of Texas at Austin Cloud Computing
McKinsey: Scaling bigger, faster, cheaper data centers with smarter designs
