Sorting Out Data Center Standards and the Mystical 4-Tier System

Sorting Out Data Center Standards and the Mystical 4-Tier System

May 1

If you run a small business, then a traditional office space that houses a remotely accessible server will likely suffice. On the other hand, if your business requires access to multiple servers and more networking equipment than the typical office can handle, then you’re probably better suited for a data center. Ideally, a data center will provide centralized storage, connectivity, and security that fully supports your hardware and daily business operations. But not all are created equal.

Data center standards are generally defined by the four-tier classification system originally created by the Uptime Institute in 1995. This system was introduced to distinguish facilities by design aspects such as building features, redundancy, and guaranteed uptime. In simple terms, it helps prospective tenants identify facilities that are best suited to meet the specific availability demands, operational needs, and compliance requirements they may have. According to the Uptime Institute, more than 500 facilities in 66 countries have been certified under these standards.

Similar systems introduced by the Telecommunications Industry Association (TIA) and Building Industry Consulting Service International (BICSI) have created some confusion over who’s got it right. We’ll try to keep things simple by examining the globally recognized criteria data center standards share across each individual tier.

Tier I

A tier 1 data center provides the bare minimum in terms of infrastructure accommodations. In addition to only having one electrical path for powering and cooling your equipment, there is little to no redundancy, meaning the failure of a single component could mean downtime. Speaking of which, tier-1 facilities offer an expected uptime of at least 99.671 percent, which works out to about 29 hours of downtime per year. That could be acceptable for a smaller business, but not so much for an enterprise where downtime costs can add up really quickly.

Tier II

Like their first tier counterparts, tier-2 data centers typically only run power and cooling distribution via a single path. However, this limitation is made up for in redundancy. For instance, these facilities often have backup generators and cooling systems that can be switched on in the event of a power outage. Boasting at least 99.741 percent uptime, tier-2 data centers promise roughly seven fewer hours of downtime, making them a wee bit more reliable than what’s available at level 1.

Tier III

Redundancy is a major point of emphasis in the tier-3 data center architecture. Built on an N+1 infrastructure, data centers at this level are equipped with everything needed to operate on a day-to-day basis plus backups of each component to be on the safe side. As a result, they guarantee at least 99.982 percent uptime (1.6 hours of downtime), which offers peace of mind for larger companies needing to conduct mission-critical operations and routine maintenance with minimal disruptions.

Tier IV

Tier 4 houses the Holy Grail of data centers. At this level you’ll find fully redundant, fully fault-tolerant environments that continually thrive in the event of multiple failures. They often run an N+2 infrastructure, so there are two extra cooling units, UPS systems, and backup generators on deck. Tier-4 data centers have at least 99.995 percent expected uptime, which works out to 0.8 hours of downtime per year – or less than half the interruptions experienced in tier-3 environments. These facilities are used by national governments, tech giants, and others that demand the maximum availability.

Calculating the Cost of Data Center Standards

Contrary to what sounds good, you don’t always need 24/7/365 availability to be successful. Though a tier-4 setup may seem like the most bulletproof standard, the most cost effective option may lie in a lower tier after working out the math. At the end of the day, weighing your specific IT needs against how much you can realistically afford to spend is the best way to find a data center solution that delivers the highest level of reliability and lowest operational cost.

Photo Credit: Michael Gray via Flickr