Go from zero to hero with our data centre cheat sheet of definitions.
Not entirely confident about your data centre know-how?
It’s completely normal to feel nervous, frustrated or in-the-dark about what data centre decisions to make and how to make them. Many companies have concerns over their data storage, safety and security, but aren’t sure where they should start to get protected. Go from zero to hero with our easy-to-understand glossary, and become confident about data centre lingo so you can be more prepared to make the best decisions for your organization.
A data centre is the brains of your operation in essence. It brings your critical IT systems together under one roof to manage and disseminate your data across multiple servers, creating a safe and secure place for it to be stored. But it’s not just a building full of servers and cables. Over the years, the data centre has evolved, using virtualization (see definition below) to optimize various resources. And, with the cloud and on-demand services now in the mix, companies have more options than ever before, from colocation to data centres as a service.
Here are the key data centre terms, minus the IT jargon.
Ambient air cooling: A data centre needs to stay cool, but A/C can cost big bucks. Ambient air cooling, or free cooling, uses naturally cool air around the data centre to reduce the heat inside. Although not technically free, savings can be substantial — plus, it’s more environmentally friendly. And in a country like Canada, this method makes a whole lot of sense in the wintertime.
Cloud computing: Everyone is using cloud apps these days, but cloud is also a platform that allows you to ‘rent’ capacity from a cloud provider. Rather than having to build your own data centre, you simply rent space on a monthly basis, increasing or decreasing your requirements as needed. For a smaller business, this can be much more cost effective. Note: If you’re subject to Canadian compliance regulations, make sure your data is being hosted within Canada.
Colocation: Think of this as having data centre roommates. This option allows you to lease a small amount of space within a data centre shared by multiple tenants. This is ideal if you only need a very small amount of space (say, 500 square feet). No need to fear: your equipment is typically separated from other tenants by rooms or cages.
Converged data centre: In a converged environment, your hardware and software come from the same vendor (everything from servers and storage, to networking, operating systems, management tools and applications). This can make it easier to manage a data centre (fewer technical skills required) or move to the private cloud, though typically only components of a data centre are converged rather than an entire data centre. Different vendors have different specialties, so a converged data centre isn’t always desirable.
Data centre as a service (DCaaS): This is pretty much what it sounds like, in the figurative sense. You rent or lease access to a third-party provider’s computing resources… and pay for that as a service. This is ideal for organizations that are tight on capital or IT staff, or have run out of room in their own data centre. DCaaS allows you to outsource even a portion of your data centre, which you can then remotely access.
Data center infrastructure management (DCIM): This is what happens when IT and building facilities converge. DCIM tools allow administrators to get a holistic view of a data centre’s performance, so they can see if energy and equipment are being used as efficiently as possible (read: cost-effective).
Powered shell: This is essentially a ‘shell’ of a data centre, without the guts. That means you can lease the facility and build out the internal components to meet your unique requirements. But, the data centre provider owns the ‘shell.’ It has several benefits, including a faster time to market.
PUE: Power Utilization Effectiveness is a unit of measurement, but rather than measuring length or width, it’s measuring the efficiency of a data centre environment. Basically, the formula takes the total facility power usage and divides it by the equipment power usage. The lower the PUE ratio, the more efficient (and cost-effective) the environment. The best data centres have PUEs around 1.5.
Resiliency: These days businesses can’t afford any downtime. Resiliency is the ability to bounce back after an equipment failure, power outage or even a security breach. This should be part of an organization’s disaster planning, and should include the use of redundant components, systems or facilities. This also includes UPS. Keep reading…
UPS: Uninterruptable Power Supply is used for power backup in data centres. In a rotary UPS configuration (think rotary phone), a continuously spinning ‘wheel’ provides backup until another source can take over, providing seamless — and yes, uninterrupted — power.
Software-defined data centre (SDDC): In this iteration of a data centre, all infrastructure (networking, storage, CPU and security) is virtualized and delivered as a service. So, rather than using hardware, the entire data centre infrastructure is provided through software.
Turnkey data centre: Similar to powered shells, these are facilities that allow you to move in your own equipment (so you lease the space, not the equipment). Unlike a powered shell, though, there’s a bit more to them: they have raised floors, power backup and fire suppression, to name but a few added perks.
Virtualization: This word keeps cropping up, and that’s because it’s so important in today’s data centre environment. Virtualization is the ability to use software to simulate hardware such as a server, storage device or a network resource. Using special software, you can maximize your existing resources, while cutting costs. There are different types of virtualization, too. Network virtualization allows you to split up bandwidth into channels, independent of each other. Storage virtualization allows you to pool physical storage from multiple storage devices and manage it all centrally. And server virtualization does the same thing, pooling all of your server resources to manage them centrally.