Discussions around 5G tend to focus on radio technologies and frequency bands. As with any innovation, the technical aspects are important, but they are a means, not an end. Furthermore, the deployment of 5G will require a tremendous expenditure of time and money. Why will operators invest their precious resources in 5G? The answer is applications and revenue. The real drivers behind the 5G rollout will be the applications that it enables, and the revenue that operators can generate from those applications. Let’s look at a typical 5G application and try to understand how it will drive network requirements.
5G Application – Connected Cars
The idea of connected cars promises safer travel, less congestion, and a more connected user experience. There are very specific requirements of the 5G infrastructure that support this vision:
- Ubiquitous, low-latency, assured connections – It is not enough for the car to be connected. To enable applications, such as congestion control, the connection must always be available. This means assurance. There must be enough low-latency bandwidth to upload and download road condition data or maps and entertainment content. These actions all require numerous cells with fat uplinks to the network.
- Dynamic service delivery – We envision numerous features that will be useful in connected cars, but we can’t predict them all. As a result, we need infrastructure that will let us quickly construct and dynamically deploy services and applications.
- Coordination with external systems – Much of the value of connected cars flows from the ability to process the data coming from the cars. For example, car manufacturers will use vehicle data analytics to understand wear and failure mechanisms. Likewise, traffic patterns will be altered based on real-time reports of road conditions. Finally, the ability to access our beloved cat videos and restaurant ratings requires connectivity to the larger internet.
In the items above, I described the required characteristics for 5G infrastructure in terms of connected cars, but they are neither unique nor limited to connected cars. Applications such as virtual reality, augmented reality, large venues, smart cities, and the internet of things all drive similar characteristics.
5G Infrastructure Requirements
Now that we understand some of the characteristics needed to support 5G, let’s look at the corresponding requirements for the supporting infrastructure.
- High bandwidth, low latency connections – 5G may be a wireless technology, but it will drive the need for ubiquitous fiber. More fiber will be needed in the access network and in the core.
- Assurance at Layers 1-3 – Dynamic and reliable services are only possible if there is effective service assurance at multiple levels, including on the fiber, at Layer 2, and at the IP layer.
- Distributed compute – 5G infrastructure will be built using the principles of the cloud. This foundation consists of standard servers, open APIs, SDN, and NFV. As a result, we will need compute elements distributed throughout the network, including in the data center, at the central office, in the cell site, and at the customer site.
- Open multi-vendor systems – Single-vendor silos are finished. We will build 5G infrastructure using components from a wide variety of vendors, enforcing competition and enabling rapid innovation.
- Integration into higher-level systems – Realizing the value of 5G for advanced services means tying the infrastructure into other systems. Such systems include SDN controllers, NFV orchestrators, existing OSS/BSS, and external systems, e.g., traffic control.
Think About the User!
As we develop new technologies, we are sometimes so involved in the specifications that we forget about what problem we are trying to solve, the impact of the technology on the broader infrastructure, and who will pay for it. Instead, we should take a holistic view to ensure that the solutions we develop are deployable in a useful and profitable manner.