Uptime and risk management – the data centre prerogative

stephane duproz

The data centre is the heart of the modern organisation. Irrespective of where it is located, on-premises, in a collocated environment or part of a hyper-scale cloud service, companies rely on it to ensure that they have access to their applications, that their data is securely stored and that should anything go wrong that they will be able to recover quickly and efficiently.

While organisations architect their IT strategies to ensure they are protected, key criteria for all companies remain the data centre itself, and how resilient its infrastructure is. All the planning in the world would have no impact if the physical infrastructure at the data centre is unable to deliver the required availability.

Stephane Duproz, CEO at Africa Data Centres, explains that data centres are typically graded according to certifications issued by the US-based Uptime Institute, with Level 3 and Level 4 being the two primary levels that organizations focus on. Data centres certified at these levels need to guarantee 99.98% and 99.995% availability respectively – equating to 1.6 hours and 20 minutes of downtime a year.

“The primary difference between the two certification levels is that a Level 3 data centre operates on an n+1 basis – where there is one standby unit for all critical systems -whereas Level 4 data centres require all systems to have an equivalent backup system – providing full redundancy.”

“The expectation from the clients, however, is that their systems are available 24/7, but the cost implications of building the infrastructure to provide full redundancy are significant and as such, the decision to make that investment is driven by the risk appetite of the companies concerned. We typically see that it’s only those organizations that would see a significant financial impact from any downtime – such as banks – who choose to make that commitment.”

As official Levels 3 and 4 installations are dependent on achieving certification, they have become shorthand for data centres that aim to meet the specific criteria, even if they have not been officially certified. On the other hand, even data centres that are certified are continually striving to improve their service levels.

Risk management strategy

“Thus data centre operators that have Level 3 and 4 facilities, albeit not officially certified, aim to ensure that the service level across both is identical, ie zero downtime. For companies the decision then becomes not one about availability but rather about risk mitigation,” says Duproz. 

With the change in the way people work, companies are having to re-examine their priorities, focusing on how to support a workforce that is no longer centrally located. This has resulted in IT teams having to re-evaluate their risk profiles, assessing what elements of their IT infrastructure are mission-critical and what the impact would be of any downtime.

“We’ve seen our clients starting to look at the benefits that a Level 4 data centre would provide, but they need it to be affordable and this is where the efforts to deliver greater uptime pay off,” he says. “Delivering this is, however, not simple. The ability to ensure that you get the required availability in a data centre environment depends heavily on the skills of the team. Downtime is much more likely to be the result of human error than of equipment failure.”

Looking to the future

While many companies are looking to the cloud as insurance in an uncertain time, Duproz points out that there will always be applications that simply are not suited to the cloud environment. Where the need for a tightly controlled environment, be it for operational, regulatory or performance issues, dictate that it operates from its own dedicated infrastructure. In these situations, the requirement for guaranteed availability will drive companies towards those facilities that ensure that operational risk is minimized, be it an enhanced Level 3 data centre of a full Level 4 installation.

Ben Kelly (James Elin)