The evolution of cellular networks

This article originally appeared in the Feb'24 magazine issue of Electronic Specifier Design – see ES's Magazine Archives for more featured publications.

What are the key considerations for Chief Technology Officers (CTOs) of IoT-enabled businesses regarding whether to upgrade to 4G LTE or leapfrog to 5G technology?

For the vast majority of IoT applications, it does not yet make sense to migrate to 5G technology. 4G networks will be around for a very long time, well into the mid-2030s even in advanced markets and much longer in many other places around the world. Moreover, 5G’s main benefit – high data rates – does not have much overlap with the requirements of most IoT applications, which tend to focus on long battery life, small device size, and low cost. When choosing connectivity technology today, there are two key considerations: timing and cost.

For 5G technology, there are two categories of chipsets available – those that support 5G eMBB (Enhanced Mobile Broadband) and 5G RedCap (Reduced Capability). As of today, neither is available at price points suitable for the majority of IoT devices, although RedCap may evolve in that direction over time.

The first question to ask is: “Does my device need high data rates?” If the answer to that question is “no”, then the next questions need to be: “When will my device go to market?” and “How long is its lifecycle?”.

There are some device categories, such as expensive equipment with lifecycles of 10+ years, where it might make sense to make the jump to 5G. But for the majority of IoT use cases, 4G remains the better option for the time being.

Though LTE-M and NB-IoT are well-suited for many IoT applications, how can the performance requirements of mid-range IoT use cases be met?

LTE Cat 1bis has emerged as a good option for mid-range IoT use cases. Data rates are the same as LTE Cat 1 (10Mb/s DL, 5Mb/s UL), but LTE Cat 1bis uses a simpler architecture and does not support receive diversity. For most IoT applications, this is perfectly acceptable, except for use cases that require voice support. IoT devices tend to transmit more data than they receive, so the removal of the second receive antenna is a compromise that has brought down cost without sacrificing much in terms of performance. Going into the late 2020s, the standard for 5G NR RedCap will see some further complexity reductions in 3GPP Rel 18 that make RedCap a suitable technology for mid-range IoT applications. The specification freeze for Rel 18 will occur in March 2024, but it will take time for chipsets to become available. We anticipate that the earliest Rel 18 devices may reach the market is in 2026 with broader adoption starting in 2028.

How does 5G NR Reduced Capability (RedCap) compare to LTE Cat 1 or Cat 4 in terms of performance and complexity?

RedCap introduces features that may be appealing to those IoT designers targeting high-performance IoT applications. For certain configurations, the RedCap peak data rate can be higher and latency lower than LTE Cat 4. In theory, RedCap supports the full spectrum of 5G FR1 and FR2 bands, which include the IOT: NETWORKS 45 mmWave spectrum. However, in practice, we expect that most deployments will take place in FR1 (sub 6GHz) bands. It is possible that mobile network operators will support RedCap, but only on a limited number of bands, at least initially. Other network requirements are not yet clear for RedCap, which could influence both performance and complexity. For example, the RedCap specification allows for one or two receive antennas and operators may either require receive diversity or determine that a single receive antenna is acceptable, as they currently do with LTE Cat 1bis. Also, the RedCap specification allows for operation in both half duplex and full duplex. The latter requires multiple duplex filters, which increases complexity, but half duplex is less bandwidth[1]efficient and therefore not all operators may allow it on their networks.

In what ways does RedCap address the specific needs of IoT applications, such as wearables, heavy equipment, and smart factories?

Consumer-focused applications like smartwatches and other wearables must support relatively high data rates in small form factors. This is not achievable with eMBB, because the complexity of the technology puts it at odds with efforts to minimise device size. At the same time, mobile operators who sell wearables are eager to add 5G subscribers to their networks, giving them an incentive to expand their 5G offering with additional devices.

Construction, mining, and agricultural equipment are often fitted with cellular connectivity for telematics, remote diagnostics, and fleet management capabilities. For these applications, medium data rates in the LTE Cat 1 range are typically sufficient. The lifespan of such equipment extends for many years, thus requiring support for the 5G transition. Guaranteeing uninterrupted equipment connection to mobile networks well into the future should be paramount for many manufacturers. Smart factories must also guarantee the longevity of the equipment on the factory floor, especially for those deploying 5G non-public networks (NPNs).

Smart factory applications are varied, including use cases such as Ethernet replacement or robotics, ideally suited for eMBB or uRLLC (Ultra-Reliable and Low-Latency Communication). But what about sensor networks, video surveillance, autonomous guided vehicles (AGVs), and industrial wearables, like smart glasses? For NPN owners to operate within a single 5G infrastructure, it is critical to also have a medium data rate technology available.