The latency of traditional cloud architectures has to keep pace with the edge when it is enhanced by 5G.

With 5G having already moved from concept to some commercial reality in many countries around the world, a realm of never-before-seen possibilities will soon open up. We are talking about an entirely new generation of applications and use-cases emerging in retail, manufacturing, gaming, as well as automotive, that will transform the way we work, live, and play. It even extends to innovative use-cases in the maritime industry, with the use of drones to improve port surveillance.

Many of these use-cases are compute-intensive and latency-sensitive, and can unlock a broad range of opportunities for businesses and consumers alike, such as enhanced end-user experience and optimized business decision-making. IHS Markit estimates that by 2035, 5G will create a US$13.2 trillion in global economic value, generating 22.3 million jobs in the 5G global value chain alone.

Can the QoE keep up?

Given that 5G networks will support as high as 100x faster data rates and 10x lower latency when compared to 4G LTE, it is unrealistic to assume traditional centralized cloud architectures can meet the Quality of Experience (QoE) expectations for newer 5G-enabled applications and use-cases.

There is hence a need to rethink our approach to ensure that content can be processed closer to end-users, humans, and machines, because this is where content is created or consumed. It will also provide for a greatly improved QoE by significantly reducing latency.

In addition, mounting privacy concerns have resulted in data localization mandates in the region, such as China, India, Indonesia, and Vietnam. This means service providers are required to host user data in local edge data centers within each country, instead of relying on a regional or international facility.

These digital ‘privacy boundaries’ also contribute to the need for a more distributed and dynamic cloud model that involves moving cloud resources to the edge. This distributed and interconnected cloud approach is therefore an ‘edge cloud’ that is scalable, efficient, and smart enough to support network slicing and intelligent automation.

The shift to Edge Cloud

My company defines the Edge Cloud as a cloud ecosystem of storage and compute assets from multiple vendors interconnected via an agile, application-aware network that can sense and adapt to the application and use-case requirements securely, and in near real-time.

Driven by a need to get content and applications closer to end-users, the Edge Cloud will be composed of data centers—scaled down when compared to more traditional facilities—and located in much closer proximity to end-users where content is both created and consumed.

The value of processing the vast amounts of data generated by these applications locally and as a result reducing backhaul traffic back to the central cloud cannot be overstated. By processing data nearer to end-users, latency is significantly reduced to effectively improve application performance.

At the same time, it allows the network to more effectively serve the large-scale analytics required to push inferences and predictions to the devices at the edge, improving application performance and enabling enterprises and users to reap the benefits of 5G use cases that demand real-time capabilities.

With the connectivity boost provided by 5G, coupled with AI and IoT, manufacturers can streamline quality assurance processes, use predictive maintenance and more precise decision-making to reach new productivity levels. A vision of truly smart critical infrastructure, connected vehicles and more, can thus be brought to life.

A hyperconnected future

In addition to concerns around real-estate and power requirements, a key challenge for edge cloud providers will be to ensure that the performance of an Edge Cloud model can rapidly scale and adapt to meet the ever-changing demands of end-users.

According to PwC, given that the global market for edge data centers is expected to nearly triple in 2024 to US$13.5bn from US$4bn in 2017, advanced networking and software capabilities should be ensured to support this data center explosion.

The dynamic nature of the Edge Cloud model requires a framework for a highly programmable and scalable infrastructure, analytics, and intelligent data-driven automation to dynamically scale both network and application cloud resources as required to meet end-user expectations in real-time.

Additionally, predictive analytics will enable networks to become more efficient and smarter, such that they are capable of anticipating demands before requests are even made.  

Understanding the need for cloud assets to be closer to the edge, combined with network and service assurance, will be critical for service providers to deliver the connectivity required for a hyper-connected future.