While many aspects of 5G have been exciting, its real-world performance has, unfortunately, not been. In fact, for a new technology whose primary promise was supposed to be blazing fast speed, it’s been downright disappointing for most people. The reason, primarily, has to do with the fact that the promised super-fast download speeds are only possible with either millimeter wave (mmWave)-equipped 5G devices and networks, but those types of signals have limited range and limited availability, or stronger offerings in the critical mid-band frequencies (2.5-3.5 GHz), which are just starting to become available. However, the slower speeds that most 5G early adopters are experiencing are also because many early 5G networks are relatively immature, from a technical perspective.

Thankfully, hope is on the horizon in the form of network enhancements that are collectively referred to as densification. As the phrase suggests, 5G network densification refers to integrating more elements into a given space. Many people presume this primarily refers to adding more cellular transmission points into a network. While that is certainly true, it’s only a portion of the answer. There’s also a great deal of work being done to increase the number/volume of signals that can be carried on wireless networks, both by widening the range of frequencies that can carry signals as well as increasing the amount of information that can be carried on any given frequency. Collectively, the goal of all these efforts is to make 5G networks faster and to reduce latency, or lag times.

The most basic form of densification involves increasing the number of cell towers. Problem is, that’s not really easy, particularly because network carriers are running into challenges with getting approval from local governments and landowners for adding new transmission points. The situation has become so challenging, in fact, that the US FCC recently had to issue a ruling clarifying the rules for 5G network infrastructure deployment. The new ruling essentially limits how much local governments can slow upgrades to existing network infrastructure, such as cell towers.

Additionally, most of the early concepts for 5G densification depended on building and installing a lot of small cells—essentially shipping box- or even shoebox-size devices that could be used to enhance the network. The problem is, most of the small cell efforts were targeted towards mmWave, and it’s clear now that those efforts (and the technology overall) are going to take much longer to widely deploy than initially expected. Not only is it difficult to get small cells installed, the costs for the equipment remain high—and the ROI isn’t as clear for many network providers as they first thought.

In addition to small cells, another important effort around network densification has to do with antennas and a technology called Massive MIMO (multi-input, multi-output). Essentially, Massive MIMO allows for a large number of antennas to send signals simultaneously, thereby reducing the number of connections per antenna, and increasing the potential speed for each connection. Many of the major carriers started to deploy MIMO antennas for 4G networks, and now work has begun on bringing even more efficient antenna designs for 5G. One interesting point worth noting is that the lower the frequency, the larger the antenna required, and the higher the frequency, the smaller the antenna. This is a big part of what enabled the possibilities of small cells for mmWave, but not for lower sub-6 5G. MIMO antennas built to be used in mid-band frequencies (in the 2.5-3.5 GHz) range, which T-Mobile now has access to via the Sprint acquisition, offer a nice compromise in that they can fit easily on existing towers.

Densification efforts are also going on in the reuse and re-assignment, as well as the new availability of radio frequency spectrum. Though many people get easily confused by spectrum issues, a simple analogy to remember is that spectrum is like lanes on a highway and signals transmitted over a wireless network are like vehicles occupying those lanes. So, any efforts that can be made to increase the number of lanes, widen lanes, or even stack lanes on top of each other, like on a multi-level highway, can all make a big difference in how many cars (or the density of the signals) can travel over them.

One of the most important issues when it comes to spectrum is that the more of it you have, the better. In addition, the location of the frequency (along the overall range of available frequencies) is also critically important. Just as having an four-lane highway running through the middle of a dense city environment is going to be a lot more valuable to most drivers than an eight-lane road that runs out in the country, so too is having the right spectrum help carry more traffic over a 5G network.

However, a critical but little understood factor is that the amount of signal traffic that can travel over a lane of a given size does not vary by frequency, just as generally speaking, the width of lanes on big city highways don’t vary much from highways out in the country. Specifically, when it comes to spectrum, you can transfer the same amount of data across a 10 MHz channel of bandwidth at 600 MHz sub-6 5G spectrum as you can across a 10 MHz channel of bandwidth at a 30 GHz mmWave spectrum. The difference is that, just as there is a lot more open land out in the country, there’s a lot more open spectrum in the mmWave frequency spectrum than there is in the lower sub-6 GHz bands.

Back to densification, what’s starting to happen is that new technologies are enabling certain frequencies to be used in different ways. In addition, some telco providers are making decisions to leverage their spectrum assets in different ways. A technology called Dynamic Spectrum Sharing (DSS) allows telcos to use some frequencies for both 4G and 5G devices simultaneously. This is important, because it lets carriers continue to support the massive number of 4G LTE users, while also allowing the small, but growing number of 5G devices get access to the network as well. After years of discussion, a few of the carriers are starting to turn on DSS, thereby giving 5G users more “lanes” on which their data can travel.

Other carriers are taking a different approach. T-Mobile, in particular, is focusing on leveraging some new low-band 600 MHz spectrum it has dedicated for 5G as well as reassigning, or “refarming” some of its frequencies to different uses as well. While not reliant on DSS, the company says it will also use the technology for certain situations. However, its primary goal is to leverage unused spectrum and free up large chunks—up to 100 MHz—in the 2.5 GHz frequency band (often called mid-band) that they acquired the rights to when they purchased Sprint. Right now, some of that mid-band spectrum is being used for 4G LTE traffic, but by moving the LTE traffic to 1.9 GHz (which both companies had previously used for 3G and 4G), they can free up large numbers of lanes for unimpeded 5G traffic. The company already turned on some of these capabilities in New York City, and early reports are that the performance improvements for 5G devices have been impressive. In fact, there are a few documented examples of download speeds of over 750 Mbps using their 5G mid-band spectrum at 2.5 GHz.

In addition to clever reuse of existing spectrum, there are a number of other network technologies starting to be deployed that can have a positive impact on network performance. One of the most important refers to standalone mode, sometimes shortened to simply SA, which many of the major carriers, including T-Mobile, have talked about turning on later this year. Essentially, SA involves updating the network infrastructure that sits in between cell towers—where RF signals from cellular-equipped devices like smartphones are received—and the cloud providers, data centers, and other computing resources where your requests are processed for things like a web site, the latest social media posts, or any other information you want to see on your smartphone.

The first versions of 5G networks—and the ones still in place—use what’s called NSA, or Non Standalone Mode, for this network infrastructure. The reason for this is expediency, as well as a healthy dose of experience. Transitioning to a new type of radio signaling format, as well as new sets of frequencies, as was required to get 5G off the ground, was a big enough task as it was. Plus, as telco providers learned when they switched from 4G to 5G, upgrading the infrastructure is a big task as well. As a result, trying to do both simultaneously inevitably led to delays. The compromised solution of 5G NSA allowed companies to continue to use the 4G LTE infrastructure, while upgrading the radio portions of the network to the new 5G radio standard. With the enablement of SA, the two-step transition process to “true” 5G becomes complete because that infrastructure also gets upgraded to fully support the 5G specification. (For more details on this transition to 5G, see “The 4G-5G Connection”.)

The real-world benefit of turning on SA mode should be improvements in response time—also called reductions in latency—as well as somewhat faster performance. It’s important to note, however, that simply flipping the switch to SA mode won’t create a huge boost in performance, as some have suggested. The problem is that all large-scale wireless networks, particularly those using 5G, are enormously complex and consist of a huge number of interconnected parts and technologies. As a result, many core network technologies and refinements by themselves create relatively modest improvements in overall network performance. The collective whole of them all working in concert is where the real magic occurs.

One important exception to this is in some of the frequency shifts mentioned earlier. Opening up new bands of frequencies for 5G networks to use, which should happen sometime in 2021 with the launch of the C-Band frequencies, as well as potentially some from CBRS (see “CBRS Vs. C-Band: Making Sense Of Mid-Band 5G” for more details on these critical new spectrum enhancements), can make a relatively immediate large-scale jump in network performance. That’s why developments in these areas are so exciting to watch—and why shifts in network quality and performance are likely to become much more apparent based on which companies have the best RF spectrum assets to use within their networks.

Finally, another interesting opportunity to leverage the increased density of 5G networks and improve performance is through application-specific types of optimizations. It turns out, for example, that if you aren’t moving around much, there are certain elements of “overhead” necessary to track you through a mobile network that can be turned off. This in turn can be used by mobile application developers to actually increase the throughput of a given connection. Similarly, there are tricks that can be used to adjust the size of packets being delivered over a network to improve performance. Neither of these capabilities come automatically—it’s going to take effort on the part of developers to enable them. For performance and latency-sensitive applications such as mobile gaming, however, it’s easy to imagine how the effort could prove to be well worth it.

The good news overall is that 5G networks are going to improve over what they offer now, both in terms of performance and responsiveness. Even better for existing 5G users is that all of these improvements will simply make most existing 5G phones faster. There are exceptions—some of the earliest 5G phones simply don’t have the hardware necessary to take care of some of these network enhancements. Thankfully, however, not a lot of those devices were sold and more current 5G phones, like Samsung’s latest Galaxy S20, as well as their low-cost A71 and A51 5G phones, will take advantage of these improvements. In some instances, they may require software updates, but the capabilities are there.

As with many things 5G, the process of densifying and improving the performance of 5G networks is going to be a relatively long, complicated road. It’s easy to forget that we went through some of these same growing pains in the early days of 4G, so we shouldn’t be surprised to experience them with 5G as well.