If you're in the market for colocation, you've probably encountered data centers pushing their “tier” number. What do these numbers mean to you?

Colocation or other IT infrastructure service providers love to talk about their “tier” numbers. If you've been seeking out these services, you've probably asked yourself: What does this number really mean? The answer, too often, is a very vague “it depends.”

When a data center is being designed or upgraded, tier numbers and similar designators are used to classify the facilities based on specific standards. The criteria vary by the organization setting the standards, but they typically quantify such things as infrastructure, capacities, functionalities and operational sustainability.

The most prominent tiering systems you're most likely to encounter are from The Uptime Institute and the Telecommunications Industry Association. (There are others, but we'll include them in a later discussion.)

The Uptime Institute

The most widely recognized and frequently referenced data center standard is the one created by The Uptime Institute. Developed in 1995, it provides a basis for comparing the uptime - also referred to as overall availability or system redundancy - between data centers.

Using a proprietary system, The Uptime Institute will certify - for a fee - that a data center's design meets its criteria for one of four tiers denoted by Roman numerals. (Other systems use Arabic numbers.) A Tier I data center offers a single, non-redundant distribution path serving IT equipment with no redundant capacity components. At the other end is a Tier IV data center, which is fully fault-tolerant and offers 2N redundant power and cooling, among other features.

The Uptime Institute doesn't publish all the evaluation criteria for its tiers, and the tier requirements are purposefully broad to allow for what the institute calls “innovation and client manufacturer and/or equipment preferences.” Compliance with a specific tier is assessed using outcome-based confirmation tests and operational impacts.

In 2013, the Uptime Institute also introduced operational sustainability standards and added gold, silver and bronze ratings. Intertwined with the four-tier system, the new ratings are awarded based on the success of data centers' operational practices and not just design standards.

TIA-942

The Telecommunication Industry Association's TIA-942 Telecommunications Infrastructure Standard for Data Centers specifies standards for data centers' cabling systems and network design. The TIA requirements are well defined, covering physical construction, electrical power, cooling, monitoring security, redundancy, maintainability and commissioning.

First published in 2005, TIA-942's tier system draws from the structured cabling work defined in TIA/EIA-568, as well as from The Uptime Institute standard. Like The Uptime Institute system, TIA-942 classifies data centers into one of four tiers or levels. (The two have since agreed to differentiate their respective benchmarking systems, with the TIA discontinuing use of the word “tier.”)

TIA-942 has been updated to address the impact of the cloud on data center infrastructure. It now covers the newer switch fabric architectures that enable data centers to provide the low-latency, high-bandwidth, any-to-any device network that cloud computing requires.

A side-by-side look

Along with their use of four tiers or levels, The Uptime Institute standard and TIA-942 share many of the same components. Although not comprehensive, the table here shows a comparison of the two systems.

Tier/Level

Uptime Institute Standard

TIA-942 Standard

I

  • Susceptible to disruptions from planned or unplanned activities
  • Single path for power and cooling distribution
  • N+0 – no redundancy
  • Includes a generator and UPS for outages and power spikes
  • Minimum of 12 hours of generator fuel
  • Requires complete shutdown for maintenance

  • Susceptible to disruptions from planned or unplanned activities
  • Single path for power and cooling distribution
  • N+0 – no redundancy
  • May or may not have a raised floor, UPS or generator
  • Annual downtime of 28.8 hours
  • Requires complete shutdown for maintenance
  • 99.671% availability

II

  • Less susceptible to disruption from planned/unplanned activity
  • Single path for power and cooling
  • N+1 components including generators, UPS, energy storage, chillers, heat rejection, pumps, cooling and fuel tanks
  • Includes UPS and generator with 12 hours of fuel
  • Redundant components can be removed for maintenance without disruption, but distribution path maintenance may require shutdown

  • Less susceptible to disruption from planned/unplanned activity
  • Single path for power and cooling
  • N+1 – includes redundant components
  • Includes raised floor, UPS and generator
  • Annual downtime of 22 hours
  • Maintenance of power path and backbone may require shutdown
  • 99.741% availability

III

  • Normal activity will not disrupt critical operations, but unplanned activity/human error may
  • Multiple distribution paths for power and cooling with one active at any one time
  • N+1 redundancy
  • All IT equipment is dual-powered or features transfer devices
  • Includes UPS and generator with 12 hours of fuel for every “N” capacity
  • Maintains full operation with any component of distribution path removed for maintenance

  • Normal activity will not disrupt critical operations, but unplanned events could still cause disruption
  • Multiple power and cooling distribution paths with one active at one time
  • N+1 redundancy
  • Annual downtime of 1.6 hours
  • Includes raised floor and ability to maintain full operation while performing maintenance on power path or backbone
  • 99.982% availability

IV

  • Normal activity does not disrupt critical operations; can experience failure of any component with no impact
  • Multiple power and cooling distribution paths that are independent, diverse and simultaneously active
  • N+1 redundancy with physical separation
  • Continuous cooling required
  • UPS and generators required with 12 hours of fuel for “N” capacity
  • Each and every component can be removed from service for maintenance without affecting critical systems

  • Normal activity does not disrupt critical operations; can experience at least one unplanned event with no impact
  • Multiple power and cooling distribution paths
  • 2(N+1) redundancy – 2 UPS each with N+1
  • Annual downtime of 0.4 hours
  • Includes raised floor and ability to maintain full operation during maintenance
  • 99.995% availability

What's the difference?

While the table makes it look as if the two systems are nearly identical, they are very different. The Uptime Institute's system seeks to be goal-oriented rather than provide a rigid technical specification for how a data center should be designed and constructed. The TIA standard, in contrast, is very specific in terms of its ratings and the associated requirements for each level of redundancy and availability.

Unlike The Uptime Institute, TIA has no formal procedure for evaluating a data center against its standard. Nor is there an official team of evaluators to enforce its standards. However, the specific requirements for each tier level are readily available, and there is no fee for their use in evaluating a data center.

Does any of it matter?

The use of tiers has helped increase awareness of what it means to achieve various levels of data center reliability. However, there's no agreement on which system is better. Everyone has an opinion, from the journalists covering the IT industry to the organizations themselves.

There is also no governing body to make sure any of tier systems are enforced. As such, the terminology around the various standards is often used rather loosely. Some data center operators even make up their own tier definitions. Unfortunately, that can lead unsuspecting customers to think the data center is certified by one of the organizations and has met the criteria for a specific tier.

The TIA does not offer certification, although various companies may conduct audits or evaluations of data centers using the TIA standards. The Uptime Institute certifies data centers, but few - particularly in the U.S. - have gone through the process. The reasons vary, but many feel the cost for certification isn't worth it, or that certifications only have value for data centers in the upper tiers of reliability. (To determine if a data center is Uptime certified, check the organization's website.)

Some companies get their data center designs certified for a certain Uptime Institute tier but do not end up building to the associated design specs. Even if they did, unless they went on to be certified for meeting the organization's operational sustainability standards, there's no guarantee they operate their facility to the standards in the design.

Question everything

Look beyond the marketing claims. Ask a data center operator for the meaning of any tier levels they cite. Are the tiers associated with an organization like The Uptime Institute or TIA, or were they created by the data center operator? If the latter, ask for a detailed explanation as to what they mean and how they can be verified.

If tier certification is important to you, include something in your contract with a data center that specifies it will maintain the certification. (Note: The Uptime Institute's certification expires after two years.) Include a penalty if the data center lets its certification lapse. Don't take a data center operator's word for it that it is certified. Get verification.

Data center must-haves

Tier numbers and classification systems provide great starting points for evaluating a data center, but don't stop there. Conduct a thorough needs and risk assessment to determine what your data center “must haves” are. How much downtime can you tolerate? Do you have compliance requirements?

Seek out data centers that have been audited to SSAE 16 standards in North America and the comparable ISAE 3402 standards for physical security and environmental controls in Europe and abroad. Those that have been audited to meet PCI DSS and HIPAA/HITECH and that have self-certified their compliance to the EU-U.S. Privacy Shield Framework also are more likely to help meet both compliance and security needs.

Ask about the data center's expansion plans. The most successful data centers fill up quickly; you don't want yours to run out of space and not be able to accommodate your company's growth or changing needs.

Remember, a higher tier number isn't necessarily better for your organization. A Tier III or IV data center would probably not be cost effective if you are just running non-critical workloads. On the other hand, if you are running production workloads that require near 100% availability, a Tier I facility may not offer the security and reliability you need.

Bottom line: Tier numbers are only a starting point. Power specifications, carrier information, maintenance regime, business continuity and disaster recovery plans and similar factors all must work together to meet your specific needs for the availability, reliability, and security of your IT assets.