A data centre is a collection of interconnected network computer servers which are accessible to users. A small data centre could be a single rack within an office, while a large corporate data centre might require a large warehouse with thousands of individual machines operating together.
The servers could be allocated to a single business, or multiple entities, and the machines might be organised as standalone machines with a single purpose or can be run together under a cloud operating system.
Data centres are grouped into tiers based on their characteristics, and when procuring space in them – whether on a cloud or server by server basis, these have implications for the way that they are accessed, the level of security and reliability they offer, and ultimately, the price that customers pay to use them for data or application hosting.
Globally, the data centre industry is governed by the Telecoms Industry Association. There is a body in the UK that fulfils a similar role – the Federation of Communication Services.
These organisations have created definitions for the different data centre tiers and provided a hosting company is able to meet the various criteria, they can market their data centres as being appropriate to certain tiers.
Traditional data centres are organised into Tier I to Tier IV. The same standards that define these can also apply to cloud services which can be described as “equivalent” to a particular standard for compliance purposes.
Data centre tiers are defined under 5 criteria which are as follows:
As you go up the list of Tiers from I to IV, there are ever more stringent requirements in place to meet the required standards.
Availability refers to the percentage of time that the operator guarantees that the data centre will be accessible and that information can be accessed. The lowest level of guaranteed availability is 99.67%. While this seems very high, it allows for almost 29 hours of downtime over the course of a year – just over one day. At the highest level Tier IV, the allowable downtime in a 12-month period is just over 0.4 hours – about 26 minutes. For web-based businesses, lost time means lost revenue.
Data centre redundancy refers to the level of support that is in place in the event of a planned outage. Redundant capacity means that if a machine fails, another machine is available to take its place. In a Tier I data centre, there is minimal redundancy – i.e. no guarantee of a backup. If capacity is available, then there may be a machine to fall back onto, but it is not a requirement. At Tier IV, you would expect multiple redundancy options which may include alternative machines in the same data centre or even a backup data centre that can be switched to in an emergency.
A reliable source of power for a data centre is essential to prevent outages. Tier I and Tier II data centres have no requirement for alternative sources of power and cooling, so in the event of a blackout in the area, all machines will shut down. With multiple sources of backup power, the risk of an outage is significantly reduced. Larger data centres will often have their own power supplies on-site in the form of generators to ensure consistent uptime in the event of wider issues.
From time to time, all servers will need to be re-booted. This may be due to software being updated or may be due to components being switched for reliability reasons. The impact of these outages is governed by the level of redundant capacity in the data centre, and access to an uninterrupted supply of power.
The nature of Tier I data centres means that there will be disruption to services in the event of either an unplanned or planned outage. Most companies will plan outages during times of low demand such as the middle of the night.
Cloud services such as Microsoft Azure and Amazon AWS work differently from a traditional data centre, as the individual servers are replaced by virtual machines that run across multiple physical servers. This usually means that redundancy is built-in, and also the impact of outages is minimised as it is rare that all machines will be impacted by an event.
For larger cloud providers, where multiple international data centres are available to act as backup, interruptions are minimised. Cloud Services such as Azure typically offer a level of service that’s equivalent to Tier III although the actual level will be dependent on the specification that you choose. For example, if you only have a single Virtual Machine and opt-out of some additional services, you may have less protection than you would with a Tier II data centre.
In addition to the standard cloud technology, it is possible to configure services to reach even higher levels of robustness. Our AWS & Azure Managed Services suite, CloudOps configures multiple virtual machines in your network to enable us to provide a level of security and reliability that’s slightly better than a Tier IV data centre at a much reduced cost.
For more information about how IG CloudOps can provide your organisation with a Tier IV+ Data Centre at a fraction of the cost you’d expect, please contact a member of our team today on 0203 697 0302, or visit our CloudOps functional overview information pages to discover more about your options or to request a quote, and we’ll be happy to help you.
CloudOps has been specifically developed to:
You might also be interested in: