It’s the phenomenon promising to be the future of processing. But what are edge data centres and how do they work?
We explore the world of edge computing for April’s edition of Digitalisation World magazine.
In the next few years, the average person will interact with a smart device or sensor 4,800 times a day. Put another way, that’s once every 18 seconds.
The modern world is dominated by interconnected devices and the power of the internet.
In our personal lives, smartphones offer a world of convenience and choice. We shop online. Browse the internet. Arrange our finances. Communicate on social media. All at the touch of a button or swipe of a screen.
And in the professional realm, machine-to-machine learning and automation take on more of a role. Tasks that for years have relied on humans are now the remit of the robots.
But all this processing takes up huge amounts of bandwidth and power.
Enterprise data centres and the cloud have become the fall-back solution to meet society’s thirst for data. But are they up to that task in the soon-to-be era of superfast 5G?
The main issue is that the ‘Internet of Things’ depends on real-time processing. Sending data to a centralised farm or cloud facility, processing it, then transporting it back to its source takes time.
It might only be a few milliseconds’ delay, but that’s enough to cause bottlenecks that could cause potentially catastrophic consequences.
Sensitive scientific research ruined in an instant or money-spinning factory production runs destined for the rubbish bin.
How Do Edge Data Centres Differ From Hyperscales Or The Cloud?
To put it simply, edge computing simply moves the processing power as near to where the data is generated in the first place. That means the factory, the office block, the shopping centre or high street.
This reduces the requirement to send information hundreds or even thousands of miles away. As a consequence, latency is low – any delay is minimal.
Handling information at the edge also reduces the volumes of expensive internet bandwidth. It enables the ‘cherry-picking’ of the most relevant data that needs processing. While it also allows data to be stored locally, which lowers the threat of it being hacked by cyber-attackers.
Micro, Modular Or Containerised – Edge Data Centres Explained
It goes without saying that it’s not feasible or financially-viable to build a standard data centre in any location.
Processing at the edge basically requires smaller, more flexible ‘centres of data’ suitable for restricted spaces or unusual environments.
That’s why its rise coincides with the concept of micro or modular data centres. These types of data farms are prefabricated and can be installed on site within a matter of weeks. They cost a fraction of the price of a standard enterprise facility.
Essentially, a micro data centre includes all the primary technology found in any enterprise data centre. So UPS systems, PDUs, cabling, air conditioning, monitoring and communications software.
But all these elements are compressed into a much smaller footprint.
Many edge data centres are actually housed in fire and weatherproof ‘upcycled’ steel shipping containers. This makes them suitable for abnormal installation settings such as a car park, a disused office, or even a building roof.
As long as the container meets Lloyd’s Register Container Certification Scheme (LRCCS) standards, these sorts of data centres offer the option of portability.
And because they’re made using modular principles, operators have the scalability to increase capacity when required simply by adding extra cabinets or even containers themselves.
What UPS Systems Do Edge Data Centres Need?
Because of the space restrictions, modular UPS are an ideal choice for micro data centres.
Modular uninterruptible power supplies deliver high power density than traditional monoblock units. They also offer operators ‘pay as you grow’ scalability and deliver exceptional efficiency.
