For many colocation buyers, the Uptime Institute’s Tier classification scheme is their first port of call when looking at the resilience of a potential data centre partner. And, if you’re a very downtime-sensitive company, you might be inclined to look for the best you can get – that is, a Tier 4 facility rather than Tier 3 or below.
What you’ll soon find in the commercial colocation market, however, is that Tier 4 data centres (particularly certified ones rather than claimed) are much, much less common than their Tier 3 counterparts – and much more expensive to boot. This blog explores why, as well as what you should consider when choosing between the two.
Get more insight like this in the Data centre services buyer's guide.
The Tier classification system, created by the Uptime Institute in the 1990s, was designed to be used by data centre owners and operators around the world to standardise and measure the resilience of their site infrastructure.
The tiers are progressive, starting at Tier 1 and ramping up to Tier 4, and a data centre must satisfy the requirements of the previous Tier before moving upward to the next one. Most colocation buyers will be interested in either Tier 3 or Tier 4 data centre space:
Tier 3: Concurrently Maintainable
This means that critical infrastructure components within the data centre can be replaced or maintained without compromising service delivery to data centre customers – as there are spare, or standby components already existing within the live environment to take the strain. The more redundancy available for such critical components, the less risk there is of any impact on the running of IT operations – either due to maintenance or loss of components due to failure. So, a Tier 3 data centre will provide protection against the majority of failures or outage by making critical components concurrently maintainable.
Tier 4: Total Fault Tolerance
The redundancy of the data centre is taken a step further here, where even the most serious of equipment or distribution path failures will not disrupt server availability. This would usually include (at least) doubling every single component up within a live environment, with every component having at least one standby that is able to operate seamlessly in the case of others being lost - this naturally has a large impact on build, operational and ongoing costs.
Does the difference matter?
As any data centre professional will tell you, there’s a big difference between the minimum requirements for Tier 3 and Tier 4 – and data centres are only ever classified according to the minimum requirements they meet. So it’s not uncommon to come across two Tier 3 data centres with very different levels of resilience, even if one of them will never quite be resilient enough to be described as Tier 4. This is why data centre Tier isn’t all that matters when it comes to avoiding downtime.
For a data centre to meet the requirements of Tier 4, it needs a standby mains power supply. And if you’re not a critical infrastructure site or government department, this can be an extremely costly investment for a comparatively small return – which is why there are so many more Tier 3 data centres out there than there are Tier 4 (and why many colocation providers are more inclined to invest in extra redundancy and power feed diversity than the Tier 4 classification – our Tier 3 Manchester data centre is an example of this).
As such, unless you have a very specific need for a standby mains power supply (and pretty deep pockets), it’s likely to be more realistic for you to look for Tier 3 data centre space than Tier 4.
Bear in mind, however, that you should be willing to look past Tier classification – which is often claimed and not certified – and at the factors that really affect resilience within the data centre owner’s control.
What to ask instead
The key is to ask the right questions to understand where a provider sits on the redundancy scale – if they’re right on the threshold and only do the bare minimum to meet the perceived Tier 3 spec, there’s more of a chance of failure and downtime.
Some of the areas you may want to explore include power feed diversity, UPS configuration (such as single parallel pair versus dual parallel pair, N+1 versus N+N and so on), cooling and management – all of which we recently discussed in another blog post: 5 questions to ask your colocation provider about resilience.
These are the kind of questions that’ll give you an idea of how a provider really protects against downtime, and help you choose one that’s right for your business. And that’s never as simple as looking at Tier 3 versus Tier 4 – you need to take the time to understand your requirements, assess the risks, and choose a facility that fits your budget as much as your wants, needs and aspirations.
Whilst the Tier classification system is a helpful starting point, it can also be something for data centres to either hide behind or be restricted by, and the less-informed buyers may not have the knowledge or experience to see beyond it.
Don’t let anything get past you when it comes to picking a colocation provider. Download our free guide below – it’ll help you ask all the right questions when choosing your next data centre.