As artificial intelligence (AI) and the Internet of Things spread, edge computing has quickly established itself as a vital bridge between the cloud and the location of whatever it is that’s creating data.
Essentially, the edge environment allows for faster and more efficient work with data, especially when processing is time-critical and the latency of hosted platforms would create problems. An autonomous car’s need to avoid a collision was the example shared by Michael Dell when he launched Dell’s IoT and edge computing division last year.
Edge can also be used by network owners to create new revenue streams. But how?
Mahdi Yahya is CEO of Ori. The company has just launched DNA, which it claims is the industry’s first public, on-demand, multi-access edge computing platform. He sets out the thinking behind the concept in our exclusive Q&A.
Internet of Business: For anyone still unfamiliar with the concept of edge computing, please provide us with your own preferred definition.
Mahdi Yahya: “Edge computing has traditionally allowed data to be processed closer to where it is created. It moves certain computing resources, or portions of an application, to the most logical edge of the cloud.
“In most cases, the cloud lives on top of a network of data centres, and the edge is the closest data centre to the source of generated data traffic.”
Or the closest computing resource. But Ori is different, you say. Your proposition is not edge computing that’s close to the cloud, but close to the user. Can you explain what that means?
“All forms of edge computing are driven by proximity to the source of data, whether it’s a user or a machine. What Ori is building is a logical extension to the cloud, exploiting computing power in network components that companies have been using for years, rather than a data centre.
“The network edge can consist of distributed technology, such as a mobile base station or a modem in our homes, which can all be connected. This can turn a network into an autonomous collection of computing power that can better serve location and environment-aware services.”
Ori has recently announced DNA, which is being billed as the industry’s first mobile edge computing platform. What does this offer networks and what does it offer developers?
“DNA is the industry’s first public, on-demand, multi-access edge computing (MEC) platform. For network owners – private and public – this can enable them to create new sources of revenue.
“It allows networks to easily and quickly onboard their distributed computing resources and make them available for developers to deploy applications that require real-time data processing. Ori sits in the middle and orchestrates the technical and financial relationship between the networks and developers.
“The on-demand business model that DNA offers, much like the cloud, allows developers to pay as they scale and gives them opportunities to access hyper-local resources that they have never had before.
“To develop future services that are intelligent and more localised, DNA forms a common meeting ground, where networks have the opportunity to recoup and balance investments in their upgraded infrastructure, while developers can simply innovate.”
Can you describe some specific areas where DNA could have a significant effect, and what that effect might be?
“While network services such as content, the IoT, and communication will immediately benefit from edge computing, we see MEC powering a broader shift in intelligence at the edge.
“We call this ‘decision-making at the edge’. This concept will have an immediate impact on machine learning and the execution of trained models in current and future applications.
“Machine learning models, used in many applications today, can be trained in the cloud but executed at the mobile edge, allowing applications to make real-time decisions in proximity to the source of data. This will have a huge impact on smart city applications, such as traffic management, disaster alerts, social care and healthcare.
“And as the network edge infrastructure continues to harness computing power, MEC will be an essential part of running immersive content applications, such as augmented reality (AR). This will allow AR devices to become smaller but more powerful, by offloading some of the device computing features to the network edge.
“By the time driverless cars become a common occurrence on our streets, MEC will be powering the car’s flow of data and the millions of data requests it generates that require immediate responses for the car to operate, and scan its environment. And also allow us to stream content and browse the Web while it drives us to our destination.”
We can’t ignore the fact that 5G is on its way. What will that add to the benefits that, you say, DNA can bring?
“With the worldwide anticipated rollout of 5G starting in earnest next year and into 2020, mobile infrastructure will undergo an important upgrade. Lower latency and higher data rates can be achieved, and computing capabilities to many components in the mobile network will be improved.
“We will then have the ability to use these devices for computing functions rather than just being a passive element that routes data to the cloud. Base stations are becoming cheaper, but more powerful, allowing for the installation of more edge points in a network.
“5G will also bring agile features to the networks, such as network slicing, which would make the deployment of third-party services over mobile networks much easier and more controlled. Slices would then be assigned to different applications, and each slice will sit on top of a group of edge nodes that a developer can utilise and pay for.”
Internet of Business says
The concept of distributed computing resources is gaining ground across the technology sector, with a shift away from the mega data centre model, to an extent, and closer to the point of need.
For the IoT that makes a lot of sense, especially where latency and locality are issues. But in some ways there are unanswered questions in this trend, the biggest of which is: who pays the electricity bill?
For example, when the DADI blockchain launched in July, it promised to slash the cost of the centralised data centre model, thanks to its members offering their MIPS in exchange for digital tokens.
But while such networks shift costs and responsibilities away from a central organisation, their members may find that their electricity and telecoms costs soar.
And while the likes of Apple, Amazon, or Google might be able to guarantee that some of their data centres are carbon neutral and sustainable, most consumers can’t make those guarantees.