1G, 2G, 3G and 4G networks incrementally enhanced the mobile experience in ways we could have only imagined. The evolution made our dumb devices smart, gave meaning to the term on-demand and turned us into a 24/7 society. Another evolution—or more likely a revolution is upon us as the industry readies to deliver on the promise of 5G (sub 1ms latency).
The Race toward 5G
The International Telecommunications Union-Radio (ITU-R) created the International Mobile Telecommunication (IMT) system on which 5G will be based. The standards organization has deemed the total download capacity for a single IoT enabled 5G mobile cell must be at least 20Gbps.
In contrast, the peak data rate for current LTE cells is about 1Gbps. The incoming 5G standard must also support up to a million connected devices per square kilometer. The standard will require carriers to have at least 100MHz of free spectrum, scaling up to 1GHz where feasible.
That said, the 5G standard (or collection of standards) is not mandated and telecom companies are fighting for leadership positions.
“The first step of 5G NR will be adding cRAN data centers to LTE housed in metro edge core clouds (or regional data centers)”
In the United States, AT&T is planning to be the first U.S. carrier to have 5G available in 12 cities by the end of this year. Asian telcos seem to have the inside track on getting 5G off the ground and into the air. Cities such as Hong Kong, Singapore, Seoul, Shanghai and Tokyo could be where large-scale adoption of 5G becomes standard starting as early as 2020. In Europe, the United Kingdom and Ireland are the best bets to get the next generation of mobile on-line.
5G Hype Versus Reality
The current hype around 5G imagines applications like immersive holograms, kinesthetic communication and haptic technology. These capabilities could equate to a world where it’s possible to interact—touch and feel— a computer-generated image or perform remote surgery.
But the speed required for such applications won’t come with the first rollouts. That rate will require high-density concentrations of cells as well as data. Many initial deployments will have maximum transmission of only up to 15Gbps at a latency of less than five milliseconds.
Speeds needed to deliver never-before-possible, latency-free applications like autonomous driving and holographic communications will only be found in small, dedicated areas. This technology is being called 5G New Radio (NR).
The main difference in 5G NR is the spectrum in which it will operate. mmWave technology used in 5G NR is defined as anything above 30 GHz (4G maxes out at 6 GHz but is usually around 2 GHz). Physics tells us that as the frequency spectrum increases so does the potential for propagation loss and sensitivity to blockage.
In layman’s terms, 5G NR communication experiences huge losses when it must go through solid objects or travel any significant distance. Millimeter waves are governed by the same physics as the rest of the radio spectrum and, as such, they have limitations related to their wavelength. The shorter the wavelength, the shorter the transmission range for a given power. This means that as the GHz go higher, so does the potential for blockage and losses.
Radio Access Networks (RAN) are necessary for 5G NR, as are many mmWave small cells very close to the user (data) in a very dense layout. This small cell densification forms a “community” that communicates with local core base stations to manage all of the network traffic and cache high-bandwidth content for delivery locally. A new generation of micro data centers will enable these local core base stations.
5G NR networks will be deployed in communities ranging in size from a small office building to a bit larger like a stadium and the maximum deployment size of a small city. Larger cities will have multiple communities for a single application like Enhanced Mobile Broadband (EMBB), plus dedicated implementations for specific applications like a smart manufacturing plant, farm or building.
The Technology of 5G NR
While 5G NR has tremendous promise, it’s a much bigger technological challenge compared to previous generations and will need many test cases and tweaks to performance to reach that promise.
5G NR will be the wireless front end of a converged network and data center architecture. Plus, the goal is to make the RAN of 5G more open and flexible, while supporting existing functions and have it massively scale to enable new applications.
The architecture will need to adapt from the current 4G base stations and metro clouds to a 5G NR distributed cloud technology of macro core clouds, metro edge core clouds and micro edge core clouds. All three versions of these core clouds will have network function orchestration and data cache, so telco and IT functions—once fiercely independent—will be forced to come together.
Implementing a distributed core cloud technology will take network function virtualization (NFV) to make it possible for operators to optimize, manage and maintain networks. NFV refers to the replacement of network functions on dedicated appliances, such as routers, load balancers and firewalls, with virtualized instances running as software on commodity hardware in small cells (very close to the user).
NFV is a key enabler of the coming 5G NR infrastructure, helping to virtualize all the various appliances in the network. In 5G, NFV will enable network slicing allowing for the creation of multiple virtual networks on shared physical infrastructure giving priority to more critical applications.
Distributed Cloud Architecture
In the 5G NR distributed cloud, compute, communication and content delivery is handled by multiple data centers but appears as a single source to the user in a world of accelerated data generation and high-bandwidth content transmission.
NFV in the context of 5G NR will be about more than merely moving functions to commodity hardware close to the user, it will be about enabling a distributed cloud computing environment that will be scalable, resilient and fault-tolerant. This distributed cloud architecture will be virtualized in a new way called cloud-based radio access networks (cRAN).
cRAN moves processing from base stations at cell sites to a group of virtualized servers running in an edge data center. It enables service providers to dynamically scale capacity and more easily deploy value-added mobile services at the network edge to improve the user experience.
But, where is that edge? Many people are talking about a massive deployment of micro edge core clouds extremely close to the user as the initial step for 5G NR networks. While this would make sense technically and philosophically to deliver the sub 1ms latency promised, network traffic coordination and orchestration with NFV and cRAN create large technical hurdles.
Hundreds of thousands of micro edge core clouds will eventually exist, but not right away. The first step of 5G NR will be adding cRAN data centers to LTE housed in metro edge core clouds (or regional data centers). In conjunction, 5G NR MIMO (multiple in, multiple out) and millimeter wave antennas will be deployed. This initial wave will support new 5G phones and deliver incremental improvements in network performance but only in the select community areas.
In 2019-2020, significant data center buildouts will occur worldwide in metro core clouds. While these will deliver performance improvements, it will not be sub 1ms latency because of their physical location. Sub 1ms latency will finally happen with the rollout of massive number of local micro edge data centers—most likely starting in 2021.