Full duplex: A fundamental radio tech breakthrough that could double throughput, alleviate the spectrum crunch

It is generally accepted that we’re fast approaching the limits of what is possible with radio waves, due to numerous technical constraints that have plagued RF-based wireless networks since their inception over 100 years ago. But what if one of those constraints — perhaps the biggest — was blown away?

The constraint of the duplex

One of the most well-known constraints in RF that it is generally impossible to transmit and receive at the same time on the same frequencies because the act of transmission creates a massive amount of interference for the receiver, preventing the receiver from “hearing” the desired signal coming from the environment.


 

This principle is the reason why most radio technologies that involve transmission and reception use some sort of half-duplex scheme, where various techniques are used to split the channel in two. With FDD (frequency division duplex), transmission and reception are separated into separate frequencies, far enough away in frequency space that any emissions bleeding over from the transmitter to the receiver would be firmly rejected. With TDD (time division duplex), transmission and reception share the same frequencies, but the operations of transmission and reception are separated by time. That is, transmission occurs for a period of time, then reception occurs for a period of time, and continues to alternate as such.

Enter Kumu Networks and Full Duplex

Kumu Networks is a two-year-old startup spun out of Stanford to commercialize research on self-interference cancellation technology. It has developed a way to allow for full duplex operation, which permits one set of frequencies to simultaneously transmit and receive signals, by utilizing self-interference cancellation. To put it simply, the effective efficiency of any RF-based technology operating in full duplex mode is doubled compared to half-duplex mode because it is now possible to use the full capacity of a single set of frequencies for both transmission and reception.
According to Joel Brand, Vice President of Product Management at Kumu Networks, the way Kumu enables full duplex operation is somewhat similar to how echo cancellation works. The receiver “listens” to determine the RF environment and uses algorithms to predict how the transmission will be altered by the environment and cancel it out from data received from the receiver. By doing so, the problem of self-interference goes away, and the receiver can “hear” the signal it is supposed to hear while the transmitter is active.

 
The major cost for this approach is increasing complexity in terms of software, which can slightly impact battery life, but Brand assured us that the approach was not only sound, but does not significantly impair battery life any more than using cellular or WiFi does today. Modems for cellular and WiFi are already quite complex and consume a lot of battery power to function, and relative to that, Kumu’s full duplex approach does not add any substantial amount of energy usage.

Applications of self-interference cancellation and full duplex

Self-interference cancellation technology and full duplex can be applied anywhere that RF technology is used today. Current efforts in using full duplex RF are centered on WiFi and LTE. WiFi and LTE TDD both operate on a single set of frequencies per channel, so the application of full duplex would allow for the removal of the temporal division aspects of the air interface without much effect to the rest of the air interface design. This would substantially improve the performance of WiFi and LTE TDD and removing many of the more difficult aspects of TDD-based air interfaces (such as temporal scheduling and interference) that impair latency and throughput.
Kumu’s self-interference cancellation technology will also help with FDD by allowing the creation of “smart” adaptive duplexers that let multiple frequency bands use one duplexer. One of the major issues with bands in radio is that every band needs its own pair of duplexers. Duplexers are the only component in the radio chain that require this, which leads to a lot of space being occupied on the phone just to lay out duplexers. Adaptive duplexers would allow consolidation of duplexers based on common frequencies.





 
Diagram showing Kumu’s RF front end in the radio chain, replacing traditional duplexers.

For example: 850MHz (ESMR+Cellular 850, UMTS/LTE bands 5/6/18/19/26, 27) and 900MHz (Cellular 900, UMTS/LTE band 8) would share a duplexer; US 700MHz (LTE bands 12/17, 13) and APT 700MHz (LTE band 28) would share a duplexer; and 1.7+2.1 GHz (AWS, UMTS/LTE bands 4/10), 1.8GHz (DCS, UMTS/LTE bands 3/9), and 2.1GHz (IMT, UMTS/LTE band 1) would share a duplexer. Without shared duplexers, nine duplexer pairs would be required to support these bands. With shared duplexers, that drops to three duplexers! The reduction in duplexers will enable even more bands to be supported.
FDD bands with smaller duplex gaps will also be possible now. Kumu’s technology makes it possible to use less than 10MHz for a gap between transmission and reception frequencies.

 And it’s not just wireless RF that full duplex offers benefits. Wireline RF technologies can benefit from the application of full duplex as well. An example of a wireline RF technology is DOCSIS, which is used to offer internet service over cable networks (using either coaxial cable of fiber-optic cable). If CableLabs (the consortium that develops and maintains the DOCSIS standard) decides to incorporate full duplex into a new version of the standard, then it will be possible to move from asymmetric FDD (where there are less uplink channels than downlink channels) to fully symmetric, high capacity connections and use less channels to offer the same capacity (or more channels to offer even more capacity and performance). Symmetric 1Gbps connections over DOCSIS would become fully feasible, allowing for an easy upgrade path for cable companies to compete with fiber internet providers like Google and C Spire Fiber. [Read: 43Tbps over a single fiber: World’s fastest network would let you download a movie in 0.2 milliseconds.]

When will it be ready?

According to Brand, Kumu Networks has already finished the first round of developing the technology and has successfully completed trials with top US and European mobile network operators. He did not reveal who they were, but given the application of the technology, it would be rather surprising if Sprint was not one of the operators in question. The finalization of the algorithms and basic designs will be completed by November. Vendors of wireless network infrastructure are working on integrating the technology (at least for reducing the complexity of remote radio heads, which over 60% of the size and weight is from industrial grade duplexers). Field trials are expected to start in early 2015.

 Miniaturization of the designs for incorporating full duplex are beginning right now, which will be necessary for bringing the technology into computers, smartphones, tablets, and other small form factor devices. The full benefit of Kumu’s technology is not possible without both the network gear and the devices supporting full duplex. Until then, the technology will largely be used to enhance existing systems.






As part of the effort to maximize the usefulness of full duplex, the IEEE (Institute of Electrical and Electronics Engineers, the group overseeing WiFi standardization development) is considering integrating full duplex capabilities into the next generation of WiFi under the High Efficiency WLAN study group. A potential consequence of incorporating full duplex into WiFi is that smaller channels can provide the same level of performance, and larger channels will provide substantially more throughput because time division duplex will be obsoleted completely. [Read: How to boost your WiFi speed by choosing the right channel.]
Similarly, NGMN (Next Generation Mobile Networks, a group chartered to design and develop proposals for next generation cellular wireless standards) is also considering full duplex for 5G cellular. If adopted, this will effectively double the amount of available spectrum for 5G cellular. Current spectrum allocated for FDD systems today would be able to support two full duplex wireless networks, rather than just the one FDD network. This would massively alleviate spectrum availability issues. It would also make it easier to identify usable low-band spectrum, since two sets of frequencies for low-band FDD would no longer be required.

 At this point in time, no vendors in RFoG (RF over Glass, used by cable companies to deliver cable service over fiber-optic lines) or RFoC (RF over Coaxial, used by cable companies to deliver cable service over coaxial lines) have indicated any interest in full duplex RF for DOCSIS. However, we feel it is a matter of time before someone starts working on it after it is incorporated into wireless.

 f all goes well, we can expect to start seeing prototype products next year that support full duplex wireless from enthusiastic vendors of network gear, such as Samsung Networks (a division of Samsung Electronics that makes cellular and WiFi network gear), Huawei (the largest network equipment vendor in the world), Cisco, and others. A particular type of product that Brand mentioned would be possible with this is LTE-Advanced small cells that reuse the same frequencies as the macro network for backhaul (network connectivity) and access (for devices to connect to) simultaneously. This allows for mobile network operators to just “plug-in” small cells and increase usable capacity in existing frequencies (according to Brand, macro networks typically operate at around 20% spectral and capacity efficiency due to the large number of users that are nearer to the edge of the cell range and cannot access the full capacity).

Conclusion

Kumu Networks has broken one of the cardinal rules of RF technology. In doing so, it has unleashed much more potential for RF-based wireless and wireline technology than was possible before. It has also created a path to effectively double the amount of capacity available from existing spectrum. Most importantly, it has created the opportunity to harmonize the next generation of global RF-based standards and make truly global wireless a reality.
Extremetech.com