Close to the edge
Edge computing is the new enterprise buzzword, but what is it really all about? Jane McCallion investigates.
For decades, IT infrastructure has remained the same. There was a central datacentre where all the corporate information was stored, which was accessed through staff's desktop PCs or, latterly, laptops and mobile devices.
Even the cloud, supposedly the very antithesis of the on-premise datacentre model, in fact operates the same way; as a centralised repository. It just happens to be rented, remote infrastructure, rather than owned by the organisation using it.
Slowly but surely, however, this structure is being turned inside out, with distributed infrastructure playing an increasingly important role.
Welcome to the Edge.
Decentralised data and connected things
When vendors and market watchers refer to "edge computing" sometimes simply referred to as "the edge" - it can be confusing. In particular, IT professionals may be tempted to conflate it with the concept of edge of the corporate network, but the two things in fact have little to do with each other.
"Increasingly, hard network edges are disappearing as the need to deal with an ever-increasing range of external entities needs to be dealt with," says Quocirca analyst and founder, Clive Longbottom. "As such, new edges have to be defined and created particularly when it comes to the Internet of Things (IoT)."
Indeed, the growth of the IoT and the new infrastructure required to support it are roundly acknowledged as the preeminent driving force behind the move to edge computing a move that is only just beginning.
Learn about the drivers and benefits of edge computing in this whitepaper from Schneider Electric.
"For us, IoT gateways are a prime example: the new edge that is being created is between a group of IoT devices and the main network, essentially air locking devices that could be a major security issue from the rest of an organisation's network," explains Longbottom.
In a March 2017 report, Forrester analysts Sophia Vargas and Richard Fichera pointed to the rapid scalability of the IoT as one of the reasons edge computing needs to be taken seriously.
"IoT represents a dramatic change in network traffic. IoT devices may initially generate only a trickle of traffic, but as they scale into the thousands or millions, the trickles will add up to a tsunami," they wrote. "The common hierarchical network will shatter dramatically under this pressure."
It's not just about IoT, though.
"IoT is part of this growth in edge, but not the only force driving growth," says Roy Illsley, principal analyst at Ovum. "5G when it comes out will almost certainly require compute and storage capabilities at the mast. IoT is just the most visible influence, and will drive edge in certain scenarios."
All of this, combined with the low-latency needs of IoT devices, means a rethink not just of how these devices connect to the corporate network, but a complete rethink of what the network looks like in relation to them including where you put your datacentre.
Small is beautiful
Whether an organisation is operating in an environment that still has a large datacentre on a central campus, or whether migration to the cloud means its server hall has been stripped back to more of a server closet, when it comes to edge computing there will be little difference in the fundamentals of what their set-up looks like.
The nature of edge computing means it often requires infrastructure that can fit in environments that might best be described as "cosy". Oil platforms, military and factory floor settings are both examples of where access to real-time data analytics, as offered by edge computing, can make a significant difference to operations. But they're both also areas where installing a traditional datacentre would be impractical, to put it mildly.
That's why micro datacentres are so important in edge computing and why vendors are starting to put more time and money into developing and marketing them.
These appliances are compact and robust enough to be placed into inhospitable environments. They're also often self contained, meaning their cooling systems, power supplies and surge protection, and so on, as well as the storage and networking components, are all already installed and configured, reducing set-up time. Thus, they lend themselves almost automatically to the type of environment where edge computing is most useful.
The edge is about more than technology, however it's about what you can do with it. Edge computing is fast becoming one of the best ways to gain actionable insight into a business
The reasons for this are quite simple: edge brings applications, data transfer, storage and compute closer together. This reduces the amount of time it takes to collect and analyse information and, therefore, to gain actionable insight from it.
The Drivers and Benefits of Edge Computing' explores how edge computing brings bandwidth-intensive content and latency-sensitive applications closer to the user or data source. Download it here.
This, in turn, allows businesses to make faster and more informed decisions, whether that's about manufacturing processes, conditions where remote units are placed, or something more global. Having data processed close to where it's created also reduces latency, which improves the user experience.
Additional benefits include reduced transmission costs and increased privacy.
Ultimately, as we move towards a world filled with AI, machine intelligence and ubiquitous connectivity, edge computing will become ever more relevant. Rather than wait to be overtaken by this curve, organisations should be looking at their business now to determine how and when edge will become relevant to them.
2022 State of the multi-cloud report
What are the biggest multi-cloud motivations for decision-makers, and what are the leading challengesFree Download
The Total Economic Impact™ of IBM robotic process automation
Cost savings and business benefits enabled by robotic process automationFree Download
Multi-cloud data integration for data leaders
A holistic data-fabric approach to multi-cloud integrationFree Download
MLOps and trustworthy AI for data leaders
A data fabric approach to MLOps and trustworthy AIFree Download