3 Ways 5G Will Drive Edge Intelligence

5G and edge compute are closely related. With the 5G network infrastructure creating a completely new layer of “fog,” 5G will allow companies to feel more secure within their own private networks. The article dives into how 5G will drive the transformation in edge intelligence.

1468
The text
Illustration: © IoT For All

In the world of the Internet of Things (IoT), two of the most popular search terms today are “edge compute” and “5G.” However, many miss the inter-relationship of the two and the opportunity 5G deployment will present for IoT adopters to make their business processes more flexible, adaptable, effective, and intelligent.

5G’s improved network reliability and durability will drive the wider deployment of closed-loop applications. With the 5G network infrastructure creating a completely new layer of “fog,” 5G will allow companies to feel more secure within their own private networks. In the same way that cloud virtualization transformed existing business systems over the past decade, the combination of network performance and edge compute capabilities will result in an “edge virtualization” that will change the way IoT operates.

5G and Edge Compute Are Tightly Intertwined

The technology and expectations of 5G networks are common technology topics. But in the midst of all the 5G hype, three key attributes of these new networks will drive the transformation in edge intelligence:

  • Ultra-dense network node deployment
  • Increased bandwidth from the higher frequency spectrum
  • Private instances enabling interoperability and choice between public and private networks

The first two attributes combine to connect 5G to edge compute. Edge compute is a system function that provides the ability to perform digital computation in devices that exist at the edge of the IoT stack. It adds digital capabilities in the same way personal computing added a new compute capacity to enterprises in the 1980s and 1990s. As the 5G network must deploy many network nodes, as many as one every 50m2, massive new amounts of edge compute are going to emerge near the edge of IoT.  Due to the increased data collected, both the physical environment of the edge, i.e., processing power in devices, and the virtual capacities, i.e., software partitioned computing machines deployed within purpose-built edge hardware like routers, device servers, terminal servers, and gateways, will evolve.

Further, the speed of the 5G network will require more and faster compute to manage the data streams. The centralized management of the infrastructure will enable cloud-like virtualization and access to a new compute layer. When a private deployment is added, this new edge compute capability will allow enterprises to implement and manage 5G as private instances. As with cloud computing infrastructure, security- and privacy-sensitive enterprises will have the ability to leverage 5G within their firewalls and policy guidelines.

The combination of these three attributes will change how cellular networks and businesses operate at the edge.

Driving Edge Intelligence

Edge intelligence is the ability of an edge deployment to adapt, deploy, and perform business processes to deliver desired outcomes. 5G will transform how businesses operate and deliver new, better outcomes by deploying new business intelligence at the edge.

This 5G-driven IoT transformation will bring three key types of edge intelligence:

1. Improved Network Reliability and Durability Will Drive Closed-loop Applications

The dense network node deployment will create a more redundant, reliable and deterministic network, as shorter transmission distances and higher bandwidth reduce latency. In addition to the IoT stack compute levels (cloud, aggregation edge, and the physical edge), 5G will deploy a new layer of edge compute, the multi-access edge compute (MEC) layer, which will reduce transmission times required to get to this cloud-like computing capacity needed for more complex, big data-driven applications.

5G will achieve low latency by providing shorter time slots at the physical radio layer, higher frequencies such as mmWave, and moving intelligence from centralized network cores to de-centralized network cores closer to the edge. The combination of lower latency and more reliable networking is going to enable closing-the-loop on business- and mission-critical applications.

Moreover, 5G will provide network and computing performance previously only available in private or proprietary wired systems, like those used in factory automation. 5G edge automation will be one of the first ways 5G will drive edge intelligence.

2. My Own Private Fog! Security-enabling Edge Deployment

Many enterprises, particularly those in industrial controls and mission-critical market segments, have been slow to adopt public cloud compute and public networks due to low confidence in both reliability and security. 5G will address these concerns, anticipated in 3GPP Release 16, by extending the private LTE network opportunities introduced in 3GPP Release 15.

Private networks based on a shared spectrum enable customers to become independent of mobile network operators. Continuous operations and automation will no longer be subject to data plan costs and overages, sensitive data will not have to traverse public infrastructure, and the network will be dedicated, not shared with other tenants. Combine this network performance with the MEC layer that comes with 5G installation, and industrial operation leaders will find themselves back in the comfort they felt with internal, wired networks and server rooms, but without the wires and servers.

3. Edge Virtualization Will Enable Massive Retrofitting of Installed Infrastructure

The functional capabilities of 5G will enable closing-the-loop on mission-critical applications and provide confidence to enterprise operational leaders that they can trust the new network and its computing capabilities. However, the more disruptive effect of these 5G attributes will be the virtualization of the network and edge compute facilitating the retrofitting of the installed base of equipment and infrastructure with edge intelligence.

Both the network and the edge compute needed to build intelligent processes will be available and software configurable—virtualized, deployed, and delivered in the same way the cloud is today for enterprise use. Service providers and manufacturers alike will find they have new options to purchase solutions and tools such as sensors, networks to connect those sensors, compute and analysis, and closed-loop controls/algorithms, to retrofit edge intelligence on their installed base. Furthermore, proprietary, purpose-built connectivity for all types of applications, including industrial production, transportation controls, and water and energy management, can be replaced or augmented with new capabilities on the 5G infrastructure.

5G is all the rage in technology trending and forecasting discussions, but many are missing the effect that 5G will have on business operations in the form of deployed edge intelligence.

Astute systems developers will combine attributes rather than focus on leveraging them individually, and those developers will find both faster and lower-cost IoT deployments. In the same way that cloud virtualization has taken artificial intelligence applications to digital enterprise operations, 5G will create virtualization capabilities that drive that same intelligence to the IoT edge.

Scott Nelson
Scott Nelson is Chief Product Officer at Digi international. For more than 25 years he has led product development and entrepreneurial business growth as both a technology and business leader. Scott is formerly CEO/CTO of Reuleaux Technology, where he helped companies in both Silicon Valley and the Twin Cities area with strategy and new business development in the Internet of Things (IoT). After beginning his career at Honeywell in the Corporate R&D center, he spent the next 15 years at Logic PD as CTO and EVP. As technology evangelist, Scott is connected to Silicon Valley and at present is a member of multiple tech start-up advisory boards and a leading start-up accelerator, the Alchemist Accelerator. He holds a Ph.D. in applied and engineering physics from Cornell University, a doctoral minor in business administration from the Samuel Johnson School of Management at Cornell University, and a B.A. degree in physics, mathematics, and computer science from St. Olaf College.