Tier rated data centres – what does it mean and why does it matter?
A few weeks ago I did a short walkthrough video of the old OzServers data centre that we were helping to decommission in East Brisbane. In that video, I talked about some of the features that are now standard in modern data centres that just didn’t exist when that facility was built in the 1980’s (before I was born!).
That, in turn, prompted a lot of discussion in the video comments, and the discussion even spilt over to my regular #AskAlek Q&A on Facebook, where someone asked me: “
Hi, Alek. I would like to know: Why is it important to have a tier rated data centre?”
. This is a really good question – a lot of people do not understand what the different tier ratings are, and how it affects the way you choose a data centre to host your equipment.
The answer is rather in depth, and to do it justice requires a whole blog post. So hold on to your hats kids, this is going to be a wild ride!
Why are data centres important?
To start with, it’s crucial you understand the vital role data centres play in our modern society. These intriguing, fascinating and mysterious places keep you connected to every aspect of daily life. From TV news, sports, weather, chatting with friends, playing games, online banking, recipe blogs, phone calls, health advice and even cat videos – all this information that you consume passes through several data centres before reaching your phone or tablet. All those data centres do a similar job, but that doesn’t mean they’re all built to the same specifications.
How are they different?
All data centres are designed to be fault tolerant and redundant, but to what degree is determined by the economic and business goals of the companies that build and maintain them. To make comparing facilities easy and to ensure standardisation – the industry has voluntarily adopted various tier rating and certification systems.
The most common system is created and maintained by a company called the Uptime Institute. They have defined what a facility requires to have to reach a particular tier and then certifies that it does meet those requirements – thus obtaining certification as a tier rated data centre.
The Uptime Institute has split the requirements up into four tiers that are ordered (helpfully) from lowest redundancy to highest redundancy:
Tier I (one)
A Tier I data centre is not fault tolerant or concurrently maintainable at all. It would be the equivalent of your office server room. If any single component fails or needs to be shut down for maintenance (UPS, Aircon, Power), then you are going to have an outage to the services within that facility. Tier I facilities are designed on a budget for small uses or where it is a backup to another larger facility.
Tier II (two)
A Tier II data centre has some degree of fault tolerance and concurrent maintainability. Most basic to midsize data centres meet or exceed Tier II design. They can usually tolerate a failure of a component (UPS, Aircon, Power) but may experience an outage if a Single Point Of Failure (SPOF) is encountered. It may be that there is only a single generator onsite and it fails to start in a power outage, or an individual aircon is unable to cope with the load of the data centres. In maintainability, you might be able to shut down a few components without impacting operations, but some big ones would require an outage.
Tier III (three)
A Tier III data centre is where you start gaining the ability for things to fail or be taken down for maintenance without any impact to the site. An air conditioner can fail, and no one would notice (if it wasn’t for the monitoring alerts!), you can take a UPS out of service without downtime, mains power outages aren’t an issue because you have redundant generators onsite ready to handle a failure of both mains and a generator failure. Tier III is the certification that most people look for when choosing a data centre. In a nutshell, you can take any part of the data centre offline without impacting your operations.
Tier IV (four)
A Tier IV rating is the highest rating currently prescribed in the system. Not only is everything redundant, but it is fault tolerant too. Any major (and minor) component can fail, and the operations of the site will keep going without a hiccup. For example, a pipe might burst on an air conditioner and flood the room, but it won’t affect anything because all plant equipment is flood proof.
As an aside: I was lucky enough recently to have a comprehensive tour of NextDC B2 which is Australias first Tier IV Constructed Data Centre. It was my first time doing a full “behind the scenes” tour of a Tier IV certified data centre, and it was a fantastic experience. I was blown away by the sheer amount of work and attention to detail that goes into meeting the level of certification. I’m so excited that Servers Australia will launch our newest PoP in NextDC B2, to expand our network and service offerings in Brisbane.
So now that you have an understanding of what the Uptime Institue Tier ratings mean, we are now ready to answer the question: “Why is it important to have my equipment in a Tier rated data centre?”
Why is it important to have my equipment in a Tier rated data centre?
Quite simply, a Tier rated and certified data centre proves that the operator of the facility “knows their stuff” when it comes to running a data centre.
Lower tier data centres are perfect for low cost/low-risk hosting, disaster recovery/backup sites, or systems that are architected to have application level high availability across multiple locations. However, businesses who are hosting mission-critical applications, and who understand the cost and impact of downtime will tend towards the higher tiers which have the most rigorous uptime requirements (and of course the extra cost that goes with it).
It’s important to remember that the Tier system, like any certification system, is just a way to identify, minimise and control risk – it can not eliminate it. Also, not all facilities subscribe to the Tier Rating system – and just because a facility isn’t formally certified, it doesn’t mean that it doesn’t “check the boxes”! Equinix is one of the largest data centre providers in the world, and they have their way of rating and certifying redundancy within their facilities.
How do I choose a data centre?
When you are choosing a data centre for your hosting needs, there are some fundamental questions that you should be asking the operators before making your decision:
- What tier was this facility built against and when was it built?
- Do you complete regular failover testing, and if so, what is your process? What was the outcome of the last test?
- Do you have independent power pathways to the racks?
- Are the power sources actually A+B?
- Where do they meet to become a Single Point of Failure (if any)?
- How much onsite fuel and water do you hold? What are your contract terms with the suppliers?
- Are your network carriers fed into the building through diverse paths?
- Is the network backbone fed from 2 sources upstream through different entry points?
- Is the building leased or owned?
- What do your security, access control, CCTV and authorised access lists entail?
- What is your fire detection and suppression strategy? What is the response time for the fire department to arrive onsite?
This may seem like an awful lot of questions, but any data centre operator worth their salt should have no problem providing all of this for you. We get asked these questions (and more) every day at Servers Australia, and we welcome this level of scrutiny from our customers.
Don’t forget, if you’re a customer (or considering becoming one), then our Solutions Team would be more than happy to take you on a complete guided tours of any of our Sydney, Brisbane and Melbourne data centres in person. To organise a tour, just contact them at any time at 1300 788 862.