Every business wants a data center that’s always up 24/7/365. But not everybody needs to pay for 100% or 5×9 availability. We’ve put together this guide to understanding the differences between data center tiers and what that means for your company.
The Basics of Data Center Tiers
The Uptime Institute categorizes data centers into 4 levels: Tier I, II, III, and IV. Each data center tier guarantees a certain level of redundancy with its equipment and infrastructure. This is the most important certification for national and international data center service providers.
- 99.67% uptime
- No more than 28.8 hours of interruption/year
- No redundancy
- 99.75% uptime
- No more than 22 hours of interruption/year
- Partial redundancy for power and cooling
- 99.98% uptime
- No more than 1.6 hours of interruption/year
- Redundancy N+1
- Availability: 99.99% uptime
- No more than 48 minutes of interruption/year
- Redundancy 2N+1 – all components redundant and capable of supporting 100% load
Breaking Down Tier Requirements
The difference between each data center tier in terms of percent uptime guarantees seems small, at less than 0.1% or 0.01% differences.
Put in terms of permitted downtime, and it’s the difference between about a day and about an hour.
With your data in a Tier 1 center, without any redundancy, several power outages at the data center might take down your servers for a few hours each time throughout the year. If you experience longer outages, the data center risks losing its certification.
Tier 3 v. Tier 4 Data Centers
You’ll see a big jump in Tier 3 and Tier 4 standards. Tier 4 requires double the resources of a Tier 3 center. With this configuration, it is possible to manage maintenance periods without affecting the continuity of service on the servers. However, these data centers will not be protected from outages in the event of serious incidents occurring on different infrastructure components.
Designed to host mission critical servers and computer systems, Tier 4 data centers are fully redundant in terms of power, cooling, network, etc. N+1 redundancy or parallel redundancy means your servers always have power. Your server runs on an uninterrupted power supply (UPS) system designed to provide consistent power, and a backup UPS is ready to power the server if the first power supply goes down.
Business Examples for Each Data Center Tier
Enterprise level – The restaurants in your small restaurant chain process credit cards and receive email reservations. Being down for a day would slow you down, but you could revert to paper for a day and catch up pretty quickly as long as you can get access to your data. Colocation in a local Tier 1 or 2 data center should suit your needs just fine.
Industry requirements – The staff at your doctor’s office or clinic always need access to client records file server and billing software. Not having access to patient data or pharmacy software would be unthinkable. A Tier 1 data center facility might not provide the redundancy, security, and HIPAA requirements you need, so maybe a Tier 2 or 3 facility is right.
Ecommerce site – You sell products online on your site and reseller sites. Your invoicing system, inventory management, and website servers need to stay up and connected for your business to make money. Being down could mean lost revenue, customers going to competitors, and shutting down production. You’re looking for Tier 3 or 4 redundancy.
Calculating Downtime Costs
Those examples give you a starting point for your data center decision. You’ll need to do the math yourself for your company.
Compare the full cost of an outage to the cost of switching to a lower or higher tier of data center. Calculate the cost of downtime in a worst case scenario.
If your network was down, consider the cost of downtime per hour.
- How much revenue and time would you lose?
- Do you think you would recover any lost sales later?
- Is there any opportunity cost or other loss considerations?
- What does it take to recover from an outage in terms of time, cost, manpower, etc.?
- How much does it cost to bump up to a higher availability?
When downtime expenses for an availability tier exceed the expenses for colocation, consider swapping to a higher availability level. Eventually you’ll reach a point where the cost of avoiding downtime is higher than the cost of the downtime.
Comparing Enterprise Data Centers
When you’re comparing data centers, make sure you’re comparing apples to apples.
According to some experts, you can build a Tier 3 data center with a more secure, reliable infrastructure than a Tier 4 facility, depending on the equipment, staff, and setup. Uptime guarantees don’t matter if your backup server crashes and you can’t restore your data, or if your customer data becomes corrupted.
And just because a data center service provider tells you their data center is Tier 4 Gold doesn’t mean it’s certified or guaranteed by the Uptime Institute. Few data centers have tier certifications from the Uptime Institute. Some data centers acquire a Tier IV design certification but fail to build to those design specs.
A certified Tier 2 data center might have better redundancy and infrastructure than one claiming to be Tier 4 without any certification. And the price of colocation in a Tier 3 data center may be the same as a Tier 2 center, so upgrading to extra redundancy in the Tier 3 is an easy call.
Our recommendation? Complete your cost-benefit analysis, requirements list, and data center research to decide the best data center for your business.