The future requires a new kind of network and a new kind of cloud.
By 2023, 5th generation mobile (5G) will make up around one-fifth of all mobile data traffic, with subscriptions forecast to reach 1 billion. This doesn’t include the massive industrial and business revenue potential of 5G, which will run into the hundreds of billions of dollars.
And it’s just the start. Looking a little farther down the road, real-time mobile connectivity is poised to pull us into an age of connected cars, advanced machine intelligence, and blended virtual and physical realities.
So how do we make this vision a reality? By building resilient and flexible networks with computing, storage and processing power distributed in just the right places at just the right times. We call this distributed cloud. It brings together the best of telecom and cloud computing and it’s the best chance for service providers to build their next big business.
Low latency unlocks applications
Cloud services are already moving out towards the edge — away from central data centers and near the data sources. We need connected things that communicate with each other in real time, such as self-driving cars confronted with fast moving road hazards. We also need connections that move data faster than our bodies can process for a seamless AR or VR experience.
To achieve the necessary low latency, we need a distributed cloud that spans the end-device, edge sites, distributed sites, central sites, and public clouds. This requires a software-defined infrastructure and network capable of managing the immense complexity of this distributed environment.
For example, a mobile AR solution might consist of a client on a device, an edge component supporting video processing, an entertainment content delivery network/caching function at a distribution site and a backend at central sites or global clouds.
Moving compute power closer to the point of data generation and consumption has the added advantage of enabling local breakout of traffic. This saves bandwidth in the network and ensures services can be provided even if the connection to the central networks breaks.
Distributed cloud and 5G
Distributed cloud does support all existing fixed, 3G/4G, and 5G use cases. However, from a revenue perspective, it will be the 5G use cases that benefit most, particularly those with demanding requirements on scalability, security, availability, low latency and bandwidth. Examples include smart manufacturing, automotive and media and content delivery, virtual reality/augmented reality, AI and data analytics, and data storage and analytics with regulatory compliance.
Let’s look at the media and entertainment industry. Today 50 percent of today’s mobile network traffic is from video. My colleagues recently forecasted this will rise by 2022 to 75 percent. Right now, we estimate that 60 to 80 percent of video traffic can be kept locally with a distributed cloud solution. Distributed cloud with computing and storage in the network allows service providers to deploy this functionality deep inside the network, allowing them to manage the traffic demand more efficiently.
Or with smart manufacturing, if you have thousands of video cameras in a factory, it’s a good idea to process video closer to the factory so that you can decrease the data volume and make use of machine learning and storage further up in the network. There are other use cases in the automotive industry in safety, traffic efficiency services, as well as preventive maintenance.
These types of use cases account for 25 percent of the $619 billion 5G business potential by 2026, figures that Ericsson identified in a recent study of 400+ use cases across 10 industry segments.
Service providers need to seize the moment
Service providers have already put a lot of work into virtualization and cloud transformations, moving from dedicated hardware to software-based network operations to increase efficiency and cut costs.
And they can easily expand their virtualized infrastructure into multiple sites to create a basic distributed cloud infrastructure. As the industry moves towards smaller computing units — containers, microservices and functions — service providers can also shorten their time to market for new services from 12 to 24 months to one to three months.
This is the best chance in more than a decade for service providers to look beyond cost cutting and to build new business. The competition is tough, as public cloud service providers are expanding their cloud capabilities to the edge, challenging service providers in service creation and introduction.
But distributed cloud means that service providers have the chance to meet this competition and seize the future. It’s time they start building the network of tomorrow today.
Lars Mårtensson, is Head of Cloud and Network Function Virtualization Infrastructure at Ericsson. This article was prepared in partnership with Ericsson.