Both Verizon and Samsung have been emphasising significantly better than expected performance from extensive trials using early 5G millimetre-wave radios. I spoke with Derek Johnston, Head of Marketing with Samsung Network Division in North America, to uncover more detail of the trials and the likely path to commercial deployment. I’ve summarised my findings below.
What’s been trialled so far?
Verizon has been trialling “pre-commercial” 5G in the 28GHz millimetre-wave band to provide a fixed broadband service for residential and small business users. I’ve used the term 5GFT (Fixed Trial) to differentiate it from the commercial 5GNR (New Radio) standard.
The radios make full use of a massive 800MHz of spectrum, featuring beam steering and an impressive 1024 element active antenna array. These are fully integrated small cell base stations, combining the baseband processing rather than being remote radio units in a Cloud RAN architecture. This reduces the technical demands and bandwidth of the fronthaul/backhaul links to each site.
Verizon has conducted trials in 11 markets (i.e. cities) using 28GHz spectrum. Every site has been connected through dedicated fibre links but has the possibility of daisy chaining a site (i.e. using another 5G Node for backhaul) in the future.
What results have been achieved?
Both Verizon and Samsung have been pleasantly surprised with the results. Specifically, Verizon has stated that it could operate non-line-of-sight (e.g. where the signal is reflected off other buildings) and provide service to skyscrapers (e.g. multi-dwelling units) with more than 20 floors.
Dense foliage has often caused problems with point-to-point microwave links, but the nature of the wide area signal propagation in built-up areas circumvents that, resulting in high levels of performance.
Data speeds of several Gbps have been achieved, dropping to several hundred Mbps at distances of 2000 feet (600 metres). Many would have thought 200 metres would be the likely reliable upper limit for high speed service – this longer distance significantly reduces the number of Small Cells required to service a given area.
Samsung sees the focus of the initial 5G solution not as a replacement to the enormous market in point-to-point millimetre-wave backhaul equipment, but instead as a new market opportunity. 5G will compete with other fixed broadband technologies, enabling operators to introduce more choice for those living with slow and poorly performing Internet broadband service to their homes.
Radio Frequency Planning
Today’s Radio Frequency Planning tools in the 28GHz band are primarily aimed for point-to-point applications and Samsung couldn’t find anything suitable for wide area 5G. So it developed its own tool that knows not just where each building is, but what material it’s made of, in order to predict absorption and reflective properties.
Samsung provides RF planning as a service to Verizon to determine the optimum location of each cell. The relatively short range of millimetre-wave means that siting and orientation of antenna is critical. Rooftops and existing cellular antenna towers are likely to be the primary candidates.
Moving onto the commercial stage
Verizon has now shifted its focus onto deployment of pre-standard 5GNR in three to five cities. The first city to be publicly named is Sacramento with a timescale to be up and running by mid 2018. The system will evolve to support standards-based 5GNR in the second half of 2018. Hardware vendors note that their current 5G network equipment can be upgraded with software updates to many of the components and hardware swaps to others in order to cope with anticipated standards revisions in the pipeline.
Since the trials, Samsung has further reduced its 5G small cell equipment by 70% in volume with a form factor just 13 inches tall (33cm) and 8 inches wide (20cm), weighing about 20 pounds (8kg). The units are completely integrated and just need power and backhaul connections. The compact size and low weight simplify ease of installation.
In January, Samsung won the contract for Verizon’s commercial solution in Sacramento, including home routers, radio access units (i.e. Small Cells) and Radio Frequency Planning services. Samsung has worked closely with the operator and city planners regarding suitable sites for permission to install equipment. Here again, the small size and low impact make it much more acceptable to city authorities from a planning perspective.
Verizon has also stated its plans to deploy 5GNR in the 39GHz band, acquired through StraightPath for $3.1 Billion last year. It now owns 95% of the 39GHz spectrum available throughout the US.
The huge capacity of each 5GNR Small Cell and relatively small coverage area should mean that most users would be adequately served. But inevitably traffic levels will grow and more capacity will need to be deployed. This could be done as for other generations, by splitting each cell into multiple sectors or by using additional frequency bands. I suspect the coverage from 39GHz will be somewhat different to 28GHz so this may not be trivial. Additional sites would need to be identified and installed as appropriate.
Verizon has indicated that it sees a potential market of around 30 million people, which given the US population of around 300 million clearly indicates this isn’t a truly nationwide opportunity. The target will be the many homes and businesses that aren’t served (or are inadequately served) in urban and sub-urban areas. This isn’t a long distance wide area solution for sparsely populated areas but could be used to serve pockets of otherwise isolated communities. The cost of installing wireless distribution should be less than laying cables, while the high performance of 5G provides substantial bandwidth and low latency.
My guess is that it will favour regions with better weather, simple geographic topology, populated in clusters and currently poorly served by broadband providers.
Using 5G at lower frequencies
5G is also being promoted for use in the more common mobile spectrum, below 6GHz. The 3GHz to 4GHz band is a particular target, but some of the existing sub-3GHz bands could also be refarmed.
In order to benefit from the differential features of 5GNR, Derek notes that you really need a contiguous block of 100MHz to make it work well. Few US operators can spare that amount of spectrum today, with Sprint probably in the best position with its massive 2.5GHz holdings.
I would think it’s much less likely that sub 6GHz bands would be used for fixed wireless service and instead kept reserved for mobile use.
Is 5G Fixed Wireless just a US phenomenon for urban/suburban areas?
The FCC reports that the US ranks as low as 23rd out of 25 countries in terms of price for 25Mbps plans, and I am sure that it is these high prices that make the US the most attractive first market for any new fixed broadband technology.
The relatively short range of 5GNR (at millimetre-wave frequencies) means it will need to compete with traditional wireline service providers in more densely populated urban/suburban areas.
The report highlights that about 7% of the US population (approx. 25 million) is without fast broadband (defined as 25Mbps down/3Mbps up), almost entirely dwelling in rural areas. However, it suggests that this issue is more likely to be resolved using satellite broadband than anything else. It would seem overkill to use a high speed/short range 5G radio connected via the limited capacity of satellite backhaul. Why not use LTE instead?
These initial trial results are quite impressive and demonstrate the potential of 5G in these very high frequency millimetre-wave frequency bands. The huge swathe of bandwidth available (800MHz) brings out the full potential of this new technology, albeit still fairly limited in range.
The first generation of commercial 5G Small Cells has shrunk dramatically in size, volume and power consumption, making them much less obtrusive and more easily deployable than the early trial equipment.
Initially, these will be used for fixed broadband service in specific markets where the business case stacks up, but will by no means offer a fully nationwide service. I’d still expect other technologies are more likely to be used for sparsely populated areas and better served markets.
In the longer term, these 5G fixed wireless Small Cells could evolve to provide a truly mobile 5G service at 28GHz, but it seems more likely that sub 6GHz bands will be used first. It would take many more 28GHz Small Cell sites to achieve a good fully mobile service than the number needed for lower frequencies.
So while 5G fixed wireless broadband looks likely to be a long term feature of the US landscape, there is quite a separate debate about 5G mobile service in the more normal sub 6GHz cellular bands. We’ll leave that topic for another time.
I've since found this 3 minute promotional video from Verizon which reports their performance results: