F1 or family car? Choosing your Data Centre Tier

F1 or family car? Choosing your Data Centre Tier

I’m sure you’d agree that you don’t get the best value out of a Formula One car by driving it along the village lanes or through the centre of the city, fun though it might be.

Written by Arif Mohamed, Freelance Tech Writer Published Tuesday, 25 July 2017 06:00

I’m sure you’d agree that you don’t get the best value out of a Formula One car by driving it along the village lanes or through the centre of the city, fun though it might be. No, a precision machine like that is best suited for the racing circuit – whether it’s Monaco, Silverstone, or Monza. That’s because the racing Mercedes or Ferrari is designed for a specific purpose. And so it is with the digital equivalent: the four-tiered Data Centre.

Data Centres are generally built to one of four specifications, suitable for differing levels of operation. The problem is, businesses are often told, or have the impression that they need a much higher level than they actually do. So, they end up being sold a Renault RS17: Renault’s latest Formula One car, with a turbocharged V6 engine. But what they really need is a Renault Espace: a family car acclaimed for being reliable and cost-effective, which does the job and features a low break-down rate.

But enough of the car metaphors. To ensure your business is underpinned by the most cost-effective Data Centre infrastructure for you, it’s worth understanding the different tiers and what they have to offer. Because, like many other things: electronic gadgets, kitchen appliances and, yes, cars, you can end up paying over the odds for capacity you really don’t need.

The Four Tier Model

In essence, there are four recognised Data Centre tiers: packing increasing levels of performance, redundancy, and, in some cases, security, among other things. One common measure for Data Centre availability comes from the Uptime Institute (a certification body) which developed a benchmarking system that defines how much uptime per year you should get from your Data Centre hardware. But there are other metrics and benchmarks available.

Knowing your availability helps IT organisations to measure things like the performance of the data centre, return on investment, and the time or financial cost to the business of any downtime.

Non-redundant Data Centres: ideal for SMEs

The first tier has a basic level of capacity and capability which makes it suitable for small businesses and less mission-critical operations such as blog hosting. Annual maintenance of Tier 1 Data Centres requires temporary shutdowns, equating on average to 28 hours per annum: acceptable for the majority of SMEs, including small shops.

In terms of the nuts and bolts, Tier 1 Data Centres provide you with a dedicated space for your IT systems, with an Uninterrupted Power Supply (UPS), cooling equipment, and an energy generator that can meet your basic backup needs in case of a system failure. However, there’s no redundancy in the components offered, hence the slightly lower availability levels.

Partially redundant for greater resilience

The next step up is a Tier 2 Data Centre. It offers partial redundancy - in other words, duplicate equipment - for power and cooling, in addition to the components that come with a Tier 1 Data Centre.

These additional power and cooling capabilities may include UPS modules, chillers, pumps and energy generators. There could also be network links and other Data Centre components, and these are sometimes enhanced versions of those offered by a Tier 1 Data Centre.

Annual maintenance requires temporary shutdowns on average equating to 22 hours per annum. The redundancy aspect means you can replace or remove some hardware equipment without interrupting power to the core computing components.

Tier 2 is often used in conjunction with another Tier 2 site to get site resilience and increase application availability across the two, with some users focusing resilience at the software level by deploying movable workloads.

Larger enterprises need more redundancy

For even greater performance and uptime, Tier 3 Data Centres add dual-powered equipment, building on the lower two levels. With the Tier 3, sites are concurrently maintainable, so service can still be provided during maintenance windows. It’s why this model is favoured by larger businesses enterprises, corporations,  global organisations and enterprise who operate mission-critical operations.  

The dual power aspect, termed “N+1”, features fault tolerance that provides high levels of power outage protection. Twelve hours is mandatory, but most sites will be provisioned for 48 hours with the ability to run longer with standby fuel contracts. Tier 3 Data Centres are concurrently maintainable. In other words, each hardware component of the IT system can be shut down for planned maintenance or repair without impacting or disrupting the overall IT operation.

However, any unplanned activity - such as a spontaneous failures of an infrastructure component, or operational errors, could still cause an outage.

Fault-tolerant Data Centres for mission-critical workloads

A Tier 4 data centre, on the other hand, is the most robust tier: fully fault-tolerant with no single points of failure. This allows for most unplanned activity to occur while still maintaining operations.

With 99.995% availability, Tier 4 Data Centres – and some Tier 3s - have a “2N+” fully redundant infrastructure. 2N+ means the infrastructure contains twice the amount of power capabilities needed to operate, plus an additional backup generator. In fact, all the components are dual-powered and fully fault-tolerant, including storage, chillers and HVAC systems (heating, ventilation and air conditioning). Consequently, any downtime event can last less than a fraction of a second before the backup kicks in. Consequently, at this level, 12 hours of protection is mandatory, but most sites will be provisioned for 48 hours with the ability to run longer with standby fuel contracts.

Strong security is another feature, with Tier 4 Data Centres – and some Tier 3s - often featuring compartmentalised security zones that are controlled by biometric access controls. As you would expect, Tier 4 Data Centres tend to be favoured by enterprises, corporations and global organisations.  These types of businesses rely on mission-critical applications, doing things like high-volume e-commerce, financial settlements, and streaming services.

Choose the Data Centre that’s right for you

Choosing to deploy your IT estate in a Data Centre is a sensible move that will enhance the operating environment, increase robustness and is more efficient and secure than using an office location. Which Data Centre is right for your organisation is the result of a combination of factors including availability and security needs. It might be that your business will work perfectly well with a lower-tier data centre, to which you can always incorporate additional levels of security for user access or data protection. 

Or you may be better off with a combination of the tiers, for example mission critical operations may require Tier 3 or above for peace of mind, with ‘back-office’ systems in Tier 1 or 2. Data Centres are rated by their ‘lowest feature’, i.e. a Tier 4 designed Data Centre with a Tier 3 component would be rated as Tier 3 despite having Tier 4 features. Primary Data Centre operators like VIRTUS continually analyse Data Centre components, regulatory and customer requirements with every element of their Data Centre designs. This enables them to deliver the most appropriate blend of the Tier features and requirements to ensure their commercial SLAs are maintained and provide customers with peace of mind and the best value. 

At the end of the day, there’s no point spending your money on a Renault RS17 if an Espace is better suited to the job. On the other hand, if your business depends on performance, an F1 racer might be what you need.