Author(s): Geoff Blaber
5G promises a lot: new spectrum, more bandwidth, lower latency, new uses, lower cost per bit, greater efficiency, the list goes on. This is a big challenge for the industry. Not only is 5G promising a lot and heightening expectations, but many of the commonly cited applications — think self-driving cars, remote surgery and more — are distant utopian visions that will need massive network depth, coverage and core network transformation. These next-generation networks offer many advantages, but they won't all be available immediately. Similarly, some of the use cases already exist and 5G stands to refine and enhance them. This is why the industry is struggling to evaluate and communicate the value of 5G. It's a function of the continued utility of LTE and depth of ambition for 5G.
Mobile edge computing is a great example of a concept that's happening now, but that could be significantly boosted by 5G. We're emerging from the cloud era to one in which computing becomes distributed. Rather than data and processing happening within the cloud, it moves closer to the user. Performance, efficiency, security, privacy, artificial intelligence and the explosion of data are all leading factors in this transformation. Indeed, we're beginning to see the broader appearance of the "wireless edge" concept, whereby intelligence doesn't just happen in the network but on the device itself. This isn't an entirely new approach, but it is gathering wider recognition as a growing variety of workloads demand local performance without dependence on the cloud.
Data is often described as the new oil, but it, too, needs to be refined. The value needs to be extracted. Sending huge quantities of data to and from the cloud is no longer desirable or practical. Think about the vast volumes of data captured by a security camera. It's logical that analysis to find relevant footage is done on the device. The same goes for autonomous driving, where decision-making needs to be instantaneous, so it can't depend on a back-and-forth connection to the cloud.
Processing of data will take place everywhere: in the cloud, at the network edge, on the device and various points in between. This illustrates the deeply symbiotic relationship of artificial intelligence and 5G. Artificial intelligence will be the mechanism that helps extract value from data, reducing demand on the network, while 5G will become the fabric that connects a raft of different devices and end points with different characteristics and requirements. An intelligent network isn't simply a matter of enabling greater throughput and lower latency. It's a network that can prioritize and adapt in real time while the devices and applications themselves become more intelligent and smarter in how they use the network.
Applications will become more resource-aware and increasingly built to adapt to a new world where computing is distributed. 5G will deliver high throughput and ultralow latency, allowing the cloud to be positioned closer to the user. This "split rendering"could be hugely disruptive once the infrastructure has been built out. For example, a highly computing-intensive virtual reality app could be processed at the network edge and streamed to the user, or used for high-definition large-scale broadcasts through content delivery networks.
This reduces dependence on resources in the device. Going back to the virtual reality example, it could enable far thinner and lighter headset designs. Although local processing will remain critical in a landscape of distributed computing, applications stand to benefit from significant supplementary power on hand. This is a major shift in the notion of the cloud to the "edge cloud". CCS Insight predicts that for performance-critical applications, leading cloud providers will seek to partner with carriers to put their clouds as close as possible to the network edge. Unsurprisingly, given carriers' complaints of marginalization at the hands of web players, this is a central component of the 5G investment thesis.
In the 4G era, cloud and mobile were the main enablers of the app-centric disruption that followed. Nonetheless, they were designed and delivered largely independently from one another, despite the app economy benefiting from both. With 5G that notion changes, but it will need a new level of collaboration across the technology ecosystem.
This article was first published by FierceWireless on 18 September 2018.