The spectrum between 3.4 GHz and 4.2 GHz is the focus of intensifying interest from regulators and operators round the world. Until recently only really useful, in wireless broadband terms, for fixed services, it is now regarded as one of the most valuable new bands for 5G. Small cells have neutralized the disadvantages of its short range and poor indoor penetration, while it offers a large amount of capacity which, in many markets, is underused, or can be converted to mobile use by adapting current fixed wireless licences.
The biggest impetus for 3.5 GHz 5G has come from China, and this has led many other regions to prioritize this area of spectrum as a pioneer band for 5G, even though official global allocations will need to wait until the World Radio Conference in 2019. In the meantime, auctions have started in countries like Czech Republic, and there are advanced trials in China, South Korea and others. This spectrum is far closer to existing mobile bands like 2.6 GHz, and so will not require the same R&D leaps as millimeter wave frequencies like 26 GHz, which are the other great hope for significant new capacity for 5G, in bands that are currently lightly used, and could encourage new, flexible licensing schemes. For instance, the midrange bands do not raise the challenge, which is significant in mmWave, of minimizing the increased signal path loss associated with a high quality signal.
Though range and coverage usually decrease, at constant power levels, as you move up the spectrum, ZTE is claiming that with the 5G New Radio (NR), coverage can be better in 3.5 GHz than it is in 2.6 GHz using FDD-LTE. “As mobile operators prepare for 5G services, network coverage is the top challenge,” said the firm’s chief scientist, Xiang Jiying, speaking at the recent Mobile World Congress Shanghai. “ZTE believes that lower frequency 5G NR can absolutely achieve equivalent or better network coverage than 4G due to the link gain generated by new technologies such as space division multiple access (SDMA), beamforming, and terminal dual-transmission channel pre-coding.”
He continued: “Especially in densely populated urban areas, whether in non-line-of-sight (NLOS) propagation scenarios or LOS propagation scenarios, the coverage of 3.5 GHz 5G NR is superior to that of 2.6 GHz FDD LTE.”
Importantly for the economics of 5G for existing MNOs, that makes “colocation of 5G and 4G completely feasible”. Colocation will be important to reduce the number of new 5G sites required (though densification will still drive new deployments), and improve early-stage capex burdens and return on investment challenges.
However, ZTE warned that 3GPP is still to resolve the issue of re-segmentation of the 4G access network architecture, which will need to be addressed to support 4G/5G coexistence. In particular, the company joins Nokia and others in insisting that an alternative will be needed to the CPRI interface which connects basebands to remote radio heads and, in a Cloud-RAN, links the shared baseband units with radio/antennas.
“If the 4G CPRI interface mechanism is still used in 5G networks, the interface bandwidth is too high, resulting in the extremely high cost of 5G bearer networks,” said Xiang. “Therefore, the re-segmentation of the 5G access network architecture becomes inevitable. So far, seven solutions have been proposed for the segmentation of 5G access networks. If the segmentation points are closer to the RF processing layer, the requirements for the transmission bandwidth and delay are higher and the system performance is better. If the segmentation points are closer to the user plane processing layer, the requirements for the transmission bandwidth and delay are lower while the radio performance loss is higher accordingly.” Xiang wants to see the distributed, virtualized and sliced networks of 5G adopting an “innovative xHaul solution that integrates the IP and optical transmission technologies to include multiple interfaces, such as the Ethernet, CPRI and eCPRI interfaces, in the same underlying optical fibre network, support flexible end-to-end logical slicing of 5G networks, and fully satisfy the bandwidth and delay requirements for the eMBB, mMTC, and uRLLC services.” Despite this endorsement, ZTE is not a member of one of the highest profile xHaul projects, Europe’s 5G PPP initiative called 5G-XHaul (Huawei is part of that effort).
Nokia has also been paying close attention to 4G/5G interworking in the 3.5 GHz band. It recently conducted a trial with Telia in Finland, using its AirScale and AirFrame platforms for a virtualized pre-5G RAN. Nokia said it achieved latency of just 1ms, in a test which controlled robots to reposition a moving ball.
The trial took place under a temporary licence in the 3.5 GHz band, which Finland expects to open up in 2018 or later. Interworking between 5G and 4G allows LTE to be used as a base for coverage while 5G supports new use cases which urgently need better data rates and lower latency, said the partners. This is supported in the first iteration of 5G NR standards, Non-Standalone, which requires an LTE anchor network.
The US, of course, has been a leader both in opening up 3.5 GHz for mobile use and pushing pre-5G tests in millimeter wave frequencies. However, as so often in the past, it has taken its own path, threatening to weaken the push for a globally harmonized band plan and device ecosystem for 5G. In mmWave, its initial efforts are in 28 GHz and 35 GHz, but most of the world is adopting 26 GHz, as that was identified in the last World Radio Conference as a candidate for 5G.
And in 3.5 GHz, the US FCC has come up with the innovative three-tiered system of access to the band, CBRS (Citizens’ Broadband Radio Service), but this is heavily geared to LTE, leading companies like T-Mobile to raise concerns that the US band will be an “orphan”, and the opportunity to use it for 5G wasted. (The US does have different issues in 3.5 GHz from most of the world as it is used for federal applications, mainly coastal radar, not for fixed wireless – hence the need for the tiered, shared spectrum system, which will set valuable precedents for other bands in other regions.)
The new FCC chairman, Ajit Pai, clearly has CBRS and 5G issues on his mind. Last week, he circulated a draft notice of inquiry on midband spectrum, asking many detailed questions about three bands, 3.7-4.2 GHz, 5.925-6.425 GHz, and 6.425-7.125 GHz. He also asked some more general questions about expanded flexible spectrum usage in the whole range from 3.7 GHz to 24 GHz.
His main interest is whether a shared spectrum framework, potentially combining licensed and unlicensed access as CBRS does, would be appropriate for other midrange bands, most particularly in 3.7-4.2 GHz, given its proximity to CBRS. He also asked whether unlicensed spectrum usage – already expanded further within the 5 GHz band under his predecessor – could be extended even more, within the 5.925-6.425 GHz area. This could be done under the FCC’s current rules governing U-NII devices, or under new provisions.
Pai is also seeking comment on the potential for more intensive fixed or flexible use of the 6.425-7.125 GHz band; and on whether auctions – including incentive auctions like the recent one in 600 MHz – could be used to increase the availability of spectrum in the whole 3.7 GHz to 24 GHz range.
The FCC will consider the midband spectrum item at its August 3 open meeting. There will be plenty of debate between the wireless and satellite communities. Last October, the Fixed Wireless Communications Coalition (FWCC) filed a petition for more spectrum in the 3.7-4.2 GHz band to be opened up for terrestrial applications, but the satellite industry responded angrily, claiming the FWCC was repeating unsupported allegations about the “supposed adverse effect of full-band, full-arc licensing of fixed-satellite satellite service earth stations on the terrestrial fixed service”.
Separately, the FCC is evaluating the feasibility of unlicensed U-NII devices sharing the 5.85-5.925 GHz band, currently dedicated to the intelligent transport system, DSRC (Dedicated Short-Range Communications).