When an airplane moves through the air, drag is an inevitable obstacle that grows as speeds increase. Similarly, radio signals travel through a challenging environment of distortions and self-interference that increase exponentially with network density and complexity.
Self-interference occurs when a transmitter jams its own receiver. Over the last century, two different methods for getting around self-interference have seen wide adoption - Frequency Division Duplex (FDD) and Time Division Duplex (TDD). In FDD, different frequency channels separated by a substantial distance in frequency are allocated to transmit and receive so that RF filters can be used to reduce self-interference. In TDD, transmit and receive happen in non-overlapping time slots across a time synchronized network so that only the transmitter OR the receiver is active at any given time. In both methods, there is substantial wastage of valuable spectrum resources.
Such waste is becoming more and more costly as evidenced by the recent $80bn C-band auction of approximately 200 MHz of spectrum in the US alone. Recent 3 GHz spectrum allocations including the C-band one might be the last sub-6 GHz band allocation we see for 5G and thus the urgency to conserve spectrum assets could not be higher. This is because sub-6 bands can propagate further and penetrate buildings.
Given the difficult RF propagation characteristics of millimeter wave bands like 28 GHz band, more easily available spectrum in such bands only help in certain cases to alleviate the spectrum crunch as Mobile Network Operators (MNO) scramble to acquire a chunk of severely limited mid-band spectrum at almost any cost further illustrates. mmWave spectrum sold for about $2 or $3 per hertz and C-band for 100x that amount, closer to $290 per hertz.
The spectrum shortage only promises to get worse as demand for wireless performance is exploding with an ever-growing set of use cases including Industrial IoT, smart cities, metaverse and autonomous cars demanding greater capacity, higher data rates and lower latency.
How is the industry responding?
It is fast dawning upon the wireless industry including cellular and WiFi that asking for new spectrum allocations are an increasingly difficult way to solve the capacity crunch. While recent allocations within the 3 GHz band for cellular and 6-7 GHz band for WiFi provide some breathing room, spectrum is a finite resource that is fast getting depleted and there is little room to continue using decades old, sub-optimal FDD and TDD methods. As a result, there is recently a serious effort underway driven by Qualcomm, Samsung and leading 5G industry heavyweights and supported by leading industry participants to lay the foundation for the evolution of duplex operation beyond FDD and TDD to Full Duplex in 3GPP Release 18 and beyond (see Figure 1).
As it turns out, high performance Self-Interference Cancellation (SIC) that enables Full Duplex was invented by Kumu Networks founders during their PhD at Stanford University about a decade ago and subsequently commercialized by Kumu.
This technology has matured to the point of commercial deployments in US Tier-1 LTE MNO networks as Full Duplex LTE Relays, the 5G industry is just at the start of a journey towards Full Duplex (see Figure 2).
As the first step in this direction, Release 18 is focusing on 2 initiatives that deserve attention. The first one is called Sub-band Full Duplex (SBFD). As mentioned earlier, in TDD, all base-stations across the network are time-aligned using accurate GPS synchronization to transmit downlink from base-stations in certain time slots. User devices send Uplink data to be received by base-stations in non-overlapping time slots. This time separation between downlink and uplink helps avoid self-interference as the base-station transmitter is never transmitting when it needs to receive uplink from devices. However, not only is this wasteful of spectrum, but it also introduces higher latency especially for uplink as devices need to wait for their turn to send uplink. This wait can be in the order of multiple milliseconds depending on the configuration of the downlink frames and results in high latency that cannot be tolerated by XR, metaverse and even factory automation applications.
Release 18 Sub-band Full Duplex makes it possible for this uplink latency issue to be solved. As shown in the Figure 3, every slot now has downlink and uplink at the base-station.
This means devices do not have to wait their turn to send packets to the base-station. But this is only possible if there is either frequency separation or self-interference cancellation or both, at least at the base-station. 3GPP is taking a phased approach, with Release 18, requiring only the base-station to operate Full Duplex while devices remain Half Duplex. In this way, the current installed base of devices can start realizing some of the benefits of Full Duplex since no changes are needed at the device side in the case of SBFD.
Beyond SBFD, as future 3GPP releases look to harness the full power of Full Duplex, Self-Interference Cancellation will likely play an even bigger role. Although SBFD and the benefits of using Self Interference Cancellation is getting a lot of attention because of 5G-Advanced and Release 18, the company’s current focus includes enhancing 5G coverage and capacity using new deployment topologies enabled by Self-Interference Cancellation. In the case of 4G LTE, the vast majority of users of 4G LTE today are covered directly from macro base-stations deployed on cell towers, though there are a limited number of LTE repeaters and relays that augment coverage. 3GPP Release 16 already standardized 5G Relays also known as Integrated Access & Backhaul (IAB). Release 18 enhances this definition to Mobile IAB’s and Vehicle Mounted Relays (VMR) by defining single hop in-band, out-of-band backhauling, device handover and dual connectivity. A Full Duplex 5G IAB using Self-Interference Cancellation technology can be deployed in a matter of minutes as there is no backhaul to provision and no spectrum to allocate for backhaul since the access spectrum channel is reused for backhaul.
Repeaters have not been heavily favored by MNO’s in 3G and LTE in part because of lack of manageability and integration into the overall MNO network. Smart repeaters as defined in Release 18 address this issue by defining a way for MNO’s to manage Smart Repeaters using a side-band control channel. Such repeaters not only allow for single-hop operation that is transparent to the device, they also allow identification, authorization, transmit power control, beamforming control, timing and TDD configuration control and on/off functions via the control channel. Using Self-Interference Cancellation technology, Smarter Repeaters become possible with up to 1000x (~30 dB) more gain than standard repeaters and thus can enable multi-hop operation as opposed to being limited to only a single hop (see Figure 4).
Self-Interference Cancellation technology not only enables Full Duplex that doubles the value of spectrum, it also enables low latency uplink and multi-hop repeaters and mesh networks. With 3GPP finally recognizing these benefits and including it in the Release 18 standard, it is now nearly certain that the access network is going Full Duplex making Self-Interference Cancellation technology a critical piece of the puzzle.