Edge computing: where are we now and what comes next?

Edge computing is already making a difference in several industries, and more are coming on board. As it becomes more commercialized, it is also evolving, making it easier for organizations to consider adopting an edge strategy.

Simply put, the edge is where data is generated, such as a device in a factory. The edge is everywhere.

Edge computing, then, refers to processing that data close to where it is being generated, which means you can make real-time or near-real-time decisions based on that data. This can help to improve data security, lower the cost of cloud computing and address privacy challenges related to the General Data Protection Regulation (GDPR) and similar, country-specific data regulations.

Operational technology (OT) in manufacturing – the hardware and software used to monitor physical assets as well as events – has long been isolated from the evolution of cloud computing. Now, we’re bringing the concept of cloud computing into the OT environment as the Internet of Things (IoT), the ecosystem of functional connected devices, expands.

Edge doesn’t need private 5G to succeed

There is some confusion out there about edge computing being relevant only in the context of private 5G or private LTE. This is not the case. They’re complementary technologies, and edge computing works with any means of connectivity.

Its adoption was, however, fired up by the arrival of private 5G. Although this enables the transmission of high volumes of data at speed, it’s logical and more cost-effective to process much of that data at the edge.

Edge computing is one of the drivers of the convergence of IT and OT, especially in situations where it makes sense to pursue the cost savings inherent in this approach. If you have a lot of data being generated in a factory or a warehouse and you carry all that data into the cloud, it is going to be expensive.

Also, if your factory relies entirely on cloud connectivity and your internet connection fails, it will grind to a halt. But with edge computing in place, you can keep running your factory to a large degree even without an active internet connection.

Industry take-up

The manufacturing industry is taking the lead in adopting this new technology. Manufacturers have found multiple new use cases for it, including computer vision, which holds benefits for automated stocktaking, quality checks and more – although uptake can be slow, depending on the cost.

The logistics and warehousing industries are also looking at edge computing because so much automation is already happening there. Amazon's huge warehouses are fully automated, for example.

There’s been a great deal of investment in this regard in manufacturing in the UK and Europe. And although we’re seeing rising interest in the US, Asia Pacific, the Middle East and Africa, many manufacturers in these regions are simply lagging the wave of innovation.

3 areas where edge strategies could be held back

When organizations do decide to embark on an edge strategy, it is typically the lack of readiness of their applications that holds them back. These applications have to be enabled for distributed computing. And while real-time decision-making takes place at the edge, some of the decision-making may still happen in the cloud, so it’s important to get the distribution of your applications across platforms right.

A lack of skills may be another hurdle to overcome, although many organizations, particularly in manufacturing, already outsource their IT support. For them, the required skill set is less of an issue than getting their applications built or adapted to work on the edge.

A third consideration is that there has always been an air gap between OT and IT. Now, however, IT terminology and concepts, such as cloud, are entering the OT environment. It can be challenging for OT engineers to adopt these new concepts and technologies after being used to the security and rationality of the OT world for so long.

Making a start on the edge

Many organizations don’t know how to make edge computing work for them.

At NTT, we start by defining the problems they're trying to solve. Rather than selling our clients a solution and then finding a problem to solve, we first identify the problems and then find the solutions to match, no matter what technology is involved. For example, we sell private 5G, but if the solution needs Wi-Fi, that’s what we’ll recommend.

Then we help our clients build their edge ecosystem of application providers, devices and independent software vendors. It’s nearly always a multiapplication solution.

Once our clients have made a start in this area, we remain by their side, offering both edge and private 5G as a managed service. Clients get tremendous value from our managing the solutions we find for them, whether that’s through monitoring, security or software updates that we provide as a service. But those that are savvy enough to handle some of these elements on their own can still do that, too.

Keeping an eye on the bottom line

Edge computing is also becoming more cost-effective as adoption grows. There’s a return on organizations’ investment on the costs of security and traffic to the cloud alone. If you take a feed from a video camera, for example, only 5% of that feed may be actionable – and that is what you process at the edge. Otherwise, in existing models, you’d carry 100% of that video feed to the cloud and then process all of it to come up with the 5% you need. That’s expensive.

These cost considerations also limit the kind of video you can carry. Today's video feed implementations max out at a 720-pixel (p) display resolution. We are working toward 1080p and even 4K cameras that will generate a lot more traffic.

Now the cost of edge processing starts making a lot of sense – it pays for itself. You just need to design your applications and your network beautifully to recover your costs as fast as possible.

The evolution: what comes next?

Looking at what comes next for edge at NTT, we are moving toward on-site edge computing as well as near-edge, which involves having a data center close to a facility like a factory or a warehouse. We are also working on the concept of the micro data center, an all-inclusive on-premises container functioning as a data center.

The technology is evolving to the point where edge data centers or edge computing in different locations – on-premises or near-edge – will be connected seamlessly. This form of smart networking will allow you to access any resource from any site.

In retail logistics, for example, placing every application on the edge is complicated, as there are so many of them. Smart networking lets you create a type of homogeneous overlay that will allow any resource access to any application, whether it's on-premises, in the cloud or at a near-edge data center.

To make edge computing and smart networking a reality for your organization, you need a smart partner. Read more about our edge computing services.

Arvind Aggarwal is Senior Manager, Enterprise 5G Solutions and Service Development at NTT.