Subscribe

New generation of data centre management

Distributed data centres are becoming even more critical to support day-to-day operations for many companies, says George Senzere, solutions engineering manager at Schneider Electric.


Johannesburg, 03 Oct 2018
George Senzere, Solutions Engineering Manager, Schneider Electric.
George Senzere, Solutions Engineering Manager, Schneider Electric.

The data centre paradigm is evolving at a staggering rate. Even as many companies appear to be offloading their data centres in favour of pure services or platforms, data centres continue to be the foundation to those capabilities. More so, the best practice among modern companies is to opt for a hybrid strategy, where needs are met with a blend of public cloud and on-premises data centres.

"There is no single fit that works for everyone," says George Senzere, Solutions Engineering Manager at Schneider Electric. "We're seeing many different approaches to what companies need from data centres and how they access those. For example, we're seeing a rise in edge data centres as on-site data management and decisions become more useful. This is driving the need for better data centre management tools because companies have matured and managing these numerous edge data centres becomes more involved.

"These distributed data centres are becoming even more critical to support the day-to-day operations for many organisations. They aren't using just offloading as a cost-saving measure. They are now looking for efficiencies within their data centre environments."

In other words, data centres are a more dynamic part of today's business strategy. Control is very important to get the most out of data centre performance. This has prompted a significant shift in how data centres can be managed.

Closer look at data centre infrastructure

One reason why data centres are so costly is that monitoring all their variables is time- and resource-consuming. Every data centre has countless data points reporting on metrics such as availability, temperature, airflow, humidity and security. The most effective way to manage data centre infrastructure was to do so at scale, in other words, co-location services.

But not all data centre demands can be met by shifting to a co-location model. As Senzere says, there are growing reasons for having data centre footprints onsite, such as at manufacturing plants. The cost and delay of transferring data to remote servers for crucial decisions are simply not worth it. In addition to this, even companies using co-location services still want insight and reporting for their planning purposes.

This means data centre infrastructure providers are not unaffected either. They manage massive client systems and need preemptive problem spotting, reliable reporting and steady management.

All parties struggle with the same problem: any given data centre is a mix of different vendor technologies, creating complexity that is taxing to monitor. It's also quite easy to overlook something. This is where data centre management has had to step up, and it is doing so through cloud systems.

"As we all know, cloud has made services much more accessible," Senzere explains. "This applies to data centre monitoring as well: you can use a cloud-driven monitoring service that combines scale, AI and skills to focus on the information that comes out of the infrastructure of the data centre."

What to expect

What are the defining characteristics of a data centre monitoring service? Foremost, it represents centralised monitoring of multi-vendor physical infrastructure, covering power, cooling and security. These should be unified into consistent reporting and action interfaces that are uniform, regardless of the myriad devices they manage.

Those interfaces should be accessible across a variety of devices, including monitoring stations and mobile phones. Relevant persons should be alerted when something goes wrong, or could go wrong: proactive and preemptive fault-finding is a hallmark of market-leading monitoring solutions.

"The way to look at it is that a data centre is already a very advanced space. There are already many sensors and many established ways for components to report their status. But the challenge has been to combine those insights into fast and reliable ways that don't cost a fortune. While the cost of establishing a data centre can be high, it's really the support of it that is most draining over the long run, especially if it's a data centre on a remote site."

Such an agnostic and singular approach to data centre monitoring was once not feasible. But modern cloud platforms are changing that completely. Whether a company relies on the infrastructure of others, has a hybrid strategy or cultivates edge footprints, it can have full, real-time insight into what its data centres are doing. This gives control over the dynamic data centre world, a place that will help decide the future of companies.

Share