Data Center Terminology Primer
When it comes to data centers, being No. 1 is not the best. Data centers typically are ranked in tiers, with Tier I the most basic and prone to downtime and Tier IV the most robust, redundant and functional.
Tier I – IV is a standardized methodology used to define uptime of data centers. This is useful for measuring:
- Data Center Performance
- Return On Investment
Identifying the Risks to Data Center Availability
Many of the risk elements that are addressed in the tier levels were identified through user experiences. Often these experiences reflect lessons learned the ‘hard way,’ principally through unplanned outages and system downtime. The factors historically associated with business disruptions include the following:
- Nature – This includes weather related events such as tornados, hurricanes and flooding
- Human – Human-related events external to the facility including commercial transportation accidents
- Utility – Disruptions in utilities such as electrical power interruptions
- Equipment – Failures in essential equipment
- Personnel – Operator error during normal business and maintenance activities
There are several infrastructure elements to consider beyond raw computing power and the number of fiber-optic pipes in and out of the data center. These site infrastructure features include power, cooling, emergency backup capacity and functionality, height of the raised flooring, fire suppression and both logical and physical security.
Tier I data centers have non redundant power and cooling infrastructures. Tier IV data centers have all the bells and whistles—everything needed to keep them running without ever shutting down for maintenance, no matter what happens. The question for most small to midsized enterprises is where along the continuum they want their data center to fall. As always, the trade-off is between dollars and sense, although legal requirements such as Sarbanes-Oxley will weigh in, too.
The tier levels give an idea of where along the continuum a data center might fall but it is up to management to specify just how reliable a center has to be. Can you live with a few hours downtime every month? A few minutes? None?
Tier I: A single path for power and cooling distribution, without redundant components, providing 99.671% availability
Tier II: A single path for power and cooling distribution, with redundant components, providing 99.741% availability
Tier III: Multiple active power and cooling distribution paths but only one path active, redundant components, concurrently maintainable, providing 99.982% availability
Tier IV: Multiple active power and cooling distribution paths, redundant components, fault-tolerant, providing 99.995% availability
It is unlikely that the typical Small Medium Enterprise (SME) will need Tier IV redundancy unless it is involved in a highly specialized area such as banking or health care, where regulations such as the Sarbanes-Oxley rules apply.
Any SME can look at the table and determine about where their operation falls. Unless there is a legal reason or business case to certify to a given level, being familiar with tiers is a nicety, but not a necessity.
Where Your Datacenter Falls
An SME can get some idea of where its data center fits on the overall scheme of things by noting that Tier I-class data centers first appeared in the 1960s. About a decade later, standards were raised to current Tier II levels. Tier III came about in the late 1980s and early ’90s. The first Tier IV data center was developed in 1994 as part of the United Parcel Service’s Windward Project. It was the first site to assume the availability of dual-powered computer equipment and cost UPS $50 million to build.
A typical SME with between 350 and 700 seats likely has a data center that is Tier I or Tier II. A Tier I data center is susceptible to disruption from both planned and unplanned activity.
Tier I sites have computer power distribution and cooling but may not have raised floors, UPS’s, or engine generators. The critical load on these systems is up to 100% of N. Even with a UPS or generator, they likely are single-module systems and have many single points of failure. The infrastructure should be completely shut down on an annual basis to perform preventive maintenance and repair work. Urgent situations may require more frequent shutdowns.
Tier II centers have raised floors, UPS’s, and engine generators. Their capacity design is N+1, which has a single-wired distribution path throughout. Maintenance of the critical power path and other parts of the site infrastructure still requires a processing shutdown.
In a Tier III center, most functions, including preventive and programmable maintenance, repair and replacement of components, and testing of systems, can be done without disrupting operation of hardware systems.
Tier IV is the best going, but even a fault-tolerant and concurrently maintainable Tier IV site does not meet the celebrated requirement of five nines (99.999%) uptime. The best a Tier IV site is expected to deliver over time is 99.995% reliability.
Aspects Considered in Tier Performance Evaluation Category System
- Electrical Utility Service
- Lightning Protection
- Power Backbone
- UPS Systems
- UPS Batteries
- Engine Generator
- Load Bank
- Critical Power Distribution
- Mechanical Raised Floor Cooling
- UPS Cooling
- Mechanical Plant
- Support Systems Contamination
- Fire Detection and Protection
- Physical Security
- Alarms and Monitoring
The Tier Performance Standards are an owner/user set of requirements used to clearly define expectations for the design and management of the data center to meet a prescribed level of availability. The Tier Level Classification system is the foundation used by many data center owners/users, consultants and design professionals in establishing a “design-versus performance” ranking approach to today’s data center projects. Success projects start with a clearly defined set of expectations towards meeting this tier classification level including:
- Owner objectives to meet a Tier III data center
- Expectations on electrical and mechanical infrastructure
- Expectations on IT infrastructure
- Operations and maintenance criteria
- Equipment and system maintainability expectations
- Expectations for expandability without business interruptions
- Project documentation requirements
- Commissioning including functional and integrated testing
It is much cheaper and easier to use the Tier Classification process before starting construction of a new data center. Moving valves, walls, and connection points or changing capacities and redundancies is less expensive and easier during the design process. The results in extended life and life cycle operational savings of prospective certification are profound.
Going through the process depends on your needs; you have to define your requirements. But if you need a top-quality data center, this is the way to go.