The rural challenge for 5G backhaul

Distances between 20–60 km make multi-channel systems very appealing. Long haul technology is a relevant solution, especially since it offers handling of multi-channels while minimizing the number and the size of antennas needed. This is significant for total cost of ownership (TCO), both for initial investment and for power consumption. In addition, it is important to consider spectrum availability in lower frequencies where wider channels are often not yet allowed.

Reaching 5 Gbps using 40 MHz channels requires at least 14 traffic channels, and utilization of 28 MHz channels requires 18 traffic channels. Such big systems have a large footprint, with waveguides and dehydrators. Reducing the footprint and achieving a more compact and cost-efficient system extends the usage to more locations than currently for long haul systems. So, what’s the solution?

The first and most obvious solution is to use wider channels if possible. Moving to 80 MHz channels means an 8-channel system can reach up to 6.4 Gbps and a 56 MHz system with 12 channels reaches 6.6 Gbps capacity. But if wider channels are not allowed, how can the hardware needed be reduced?

Not all rural systems need to perform over distances of 75–150 km, so it is also possible to use less than maximum output power (Pout) with good results in many cases. This enables use of carrier aggregation in long haul systems. When carrier aggregation is used, the same output power needs to be distributed over two channels and since the total bandwidth is wider than one channel the maximum Pout needs to be reduced a little. In shorthaul, this technology is starting to be used in many places with good results. Adding it to the long haul toolbox will further increase flexibility, enabling users to benefit from greater channel capacity while using half the number of radios and filters.

Let’s examine what it looks like when building a 20 km long hop without space diversity in 11 GHz using 40 MHz channels and a 0.9 m antenna.

A traditional 8+0 without carrier aggregation and full Pout available per channel provides a maximum of 3 Gbps with a guaranteed capacity of 1.7 Gbps. Carrier aggregation makes it possible to instead use a 16+0 system, which still has 8 filters and radios and can reach 5.7 Gbps at peak rate with a guaranteed capacity of 2.3 Gbps.

If using carrier aggregation the total power consumption for the 16+0 system is closer to a traditional 8+0 system as fewer transmitters are used. This also impacts the footprint and the initial hardware investment, greatly benefiting TCO.

In summary, using carrier aggregation makes it possible to increase the peak capacity by 67 percent with the same availability while reducing hardware footprint, power usage and TCO compared to expansion using traditional solutions.

Line-of-sight multiple-input multiple-output (MIMO) as used in microwave is often seen as a solution when spectrum is scarce, as it has the potential to double available spectrum capacity. However, this technology is still seldom used in long haul because optimal antenna separation becomes increasingly problematic with hop distance and lower frequencies. When using less than optimal separation, the traffic channel will be at greater risk of fading due to phase variation of the signal, which can negatively impact capacity. To avoid this impacting all channels simultaniously, a link can be created, combining some channels using space diversity (SD) with some channels using MIMO in a multi-channel system. Combining these channels using radio-link bonding produces a similar solution as multi-band, with a combination of capacities with different availabilities, which are all error-free.

This ensures guaranteed traffic can have the availability of a traditional SD link, and peak capacity can be increased by up to 50 percent, or even by up to 75 percent, compared to a traditional SD hop using the same spectrum. The example in Figure 18 shows four radio frequencies in two polarizations and two antennas. A traditional 8+0 SD system achieves 1.2 Gbps at 5’9 s and 2.8 Gbps at 3’9 s availability. If instead half the channels are used as 4x4 MIMO and the rest as 4+0 SD with radio link bonding, it provides 1.4 Gbps at 5’9 s and 3.2 Gbps at 4’9 s availability, plus an astounding 4.4 Gbps for 99.8 percent of the year. However, this requires space for near optimal antenna separation in the tower, and that is normally a limiting factor. If it is possible to achieve, peak link capacity can be increased by 50 percent while keeping a reasonably high capacity with 5’9 s availability. If one can accept only 2+0 SD and make the rest into 3x(4x4) MIMO, even more peak capacity can be reached.

The challenge of getting enough transport capacity to rural 5G sites can be overcome, but it may be at the expense of some long-standing practices. We have seen in previous Microwave Outlook articles that end-user satisfaction in a 5G mobile network is linked more to peak capacity support than to the 5’9 s availability. Utilizing carrier aggregation and line-of-sight MIMO could be an efficient way to meet those expectations.