The expansion of 3GPP technologies into unlicensed spectrum bands is the result of years of standardization efforts, with the participation of stakeholders from the mobile broadband and Wi-Fi industries. Havish Koorapaty, Master Researcher at Ericsson and Vice-Chairman in 3GPP RAN working group 1, has been involved in this effort as a rapporteur for the initial study and work items. Here, he provides his perspective on how the availability of multiple technologies on unlicensed airwaves will advance the common good.
In 2016, 3GPP (Third Generation Partnership Project) cellular technologies entered unlicensed spectrum bands for the very first time with the completion of Licensed Assisted Access (LAA) with LTE in 3GPP Release 13. This year, with 3GPP Release 16, the foundation was laid to begin deploying 5G New Radio (NR) in unlicensed spectrum (referred to as NR-U) in the license-exempt 5 GHz and 6 GHz bands. The table below summarizes the evolution of 3GPP technologies for operation in unlicensed spectrum from licensed-assisted access (LAA) with LTE, where LTE is operated on an unlicensed carrier along with a carrier in licensed spectrum, to standalone operation with NR, where an NR carrier can operate independently in unlicensed spectrum (eLAA and feLAA refer to enhanced LAA and further enhanced LAA).
The long road to this milestone has involved years of cross-industry engagement and technical discussions in standards and regulatory bodies across two fiercely independent technology domains: cellular technologies based on 3GPP standards, and Wi-Fi technology based on IEEE standards.
At this point, we should also acknowledge the elephant in the room; that, the entry of 3GPP technologies into unlicensed spectrum has widely been perceived as positioning the cellular and Wi-Fi industries in direct competition with one another. When 3GPP first entered discussions to evolve LTE to operate in unlicensed spectrum, there were concerns raised that the use of unlicensed spectrum to offload traffic for services offered by cellular operators via LTE was an unfair monetization of a common good, and that it would reduce the general availability of unlicensed spectrum for other technologies.
We have a very different perspective. Users access services on devices and networks that best serve their needs. Smartphones already supported the use of unlicensed spectrum via Wi-Fi before the evolution of 3GPP technologies to operate in unlicensed spectrum was considered. Furthermore, cellular operators themselves often own and operate Wi-Fi networks to complement their licensed spectrum. In addition, there are commercial connectivity services provided purely based on unlicensed spectrum that devices can use. Thus, unlicensed spectrum already plays a role in supporting a level of user experience over devices with subscriptions on licensed spectrum. 3GPP technologies simply provide an additional option for use of unlicensed spectrum in scenarios where it has the potential to increase access to spectrum, improve spectral efficiency and enhance the end-user experience even further.
Moreover, we see many situations where the technologies complement rather than compete with one another, thereby driving the collective technological boundaries forward. For example, many LTE-LAA deployments are outdoors in areas where Wi-Fi networks are often not used or are not reliably accessible. It should also be noted that the development of cellular and Wi-Fi technologies, each with very different roots, can be a positive influence on each other. We believe that both industries can coexist and together deliver greater value to the general population. Therefore, it is clear to us that, in the end, the introduction of 3GPP technologies to unlicensed spectrum will benefit users and help better serve the common good.
Improving data rates and spectral efficiency serves the common good
3GPP technologies operating in unlicensed spectrum can increase access to unlicensed spectrum in areas where Wi-Fi would typically be unavailable, due to more robust coverage, especially on important control channels. Coverage, as well as throughput over unlicensed spectrum, may be further enhanced when carriers operating in unlicensed spectrum are aggregated with carriers operating in licensed spectrum (referred to as carrier aggregation), where control channels can be serviced over more reliable licensed spectrum.
As an example of increased access to unlicensed spectrum, when users leave their homes or offices and move outdoors, they typically remain connected to their home or office Wi-Fi system only as far as the street outside. At this point, owing to the Wi-Fi system’s limited coverage, their device will usually switch to a cellular connection. This is an instance where we see the potential for cellular technologies in unlicensed spectrum. There are essentially significant corridors with pockets of vacant unlicensed spectrum that can be reused with 3GPP technologies to increase access to and the utility of such spectrum. This increased access can allow service providers to serve a larger user base with higher data rates.
Furthermore, even when access to Wi-Fi is available in a certain area, 3GPP technologies operating in unlicensed spectrum may provide better throughput and performance, especially when carrier aggregation with licensed spectrum is used. In such cases, these users that would otherwise have been served by Wi-Fi, can be served more efficiently, thus freeing up radio resources that can be used by devices that can only access Wi-Fi networks.
Therefore, increased access to spectrum, as well as improved spectral efficiency and coverage through 3GPP technologies, benefit all users accessing services both in licensed and unlicensed spectrum.
From the very beginning, including in workshops that pre-dated the official start of standardization activities for LTE-LAA in 3GPP, key stakeholders from the Wi-Fi industry were involved in the discussions. As standardization activities took off, 3GPP became the forum where the main technical dialogue between the cellular and Wi-Fi industries started to take place. As full-fledged members of 3GPP, many key companies from the Wi-Fi industry influenced the direction and outcome of the discussions during the development of both LTE and NR for operation in unlicensed spectrum.
As standardization of LTE-LAA commenced, the regulations pertaining to operation in unlicensed spectrum came under scrutiny. Some of the most stringent regulations were part of the harmonized standards developed by the European Telecommunications Standards Institute’s (ETSI) Broadband Radio Access Networks (BRAN) committee. It was deemed by the Wi-Fi community, that the existing requirements in the harmonized standards for non-Wi-Fi devices at the time would not lead to good coexistence between Wi-Fi and non-Wi-Fi technologies. Therefore, an effort to redefine the harmonized standards started.
Right from an early stage, the discussions in ETSI BRAN became an important complement to the discussions in 3GPP, where a lot of the engagement between the industries occurred. Key representatives from both Wi-Fi and 3GPP stakeholders discussed difficult technical issues directly. This played a critical role in shaping the current harmonized standard for the 5 GHz band in Europe, developed to be able to support the operation of multiple technologies in unlicensed spectrum.
These discussions and direct interactions in 3GPP and ETSI BRAN were further complemented through a regular exchange of liaison statements and a couple of workshops bringing together all the key stakeholders. All of this served to ensure that the interests and concerns of the Wi-Fi industry were fully discussed and understood throughout the process of adapting 3GPP technologies to operate in unlicensed spectrum.
A pursuit for the best possible technological coexistence
In unlicensed bands, devices share the wireless spectrum and one method by which this can be done is to try and detect transmissions from other nodes before transmitting. This is commonly referred to as a listen-before-talk (LBT) or channel access procedure. In other words, the device listens to see if the channel is busy and, if it is, will wait until the transmission has ended before transmitting. Channel access mechanisms determine how devices coexist with one another. Hence, it was no surprise, that this became the area of greatest interest from the perspective of coexistence between Wi-Fi and 3GPP technologies.
A fundamental design philosophy that was adopted during the initial study in Rel-13, was to align channel access mechanisms for 3GPP technologies in unlicensed bands as much as possible with Wi-Fi standards. Since Wi-Fi had already been operating successfully in unlicensed spectrum, this design philosophy was the right choice to ensure good coexistence with Wi-Fi in unlicensed spectrum. Indeed, the specifications for channel access in unlicensed spectrum have a high degree of commonality and the ETSI BRAN harmonized standard mostly resembles the channel access mechanism already used for Wi-Fi. While alignment with Wi-Fi was achieved on these aspects, fundamental differences in the technologies meant that alignment was not achieved on a key aspect: the energy detection threshold used for deferring to neighboring devices.
The simplest method for detection of signals during the LBT procedure is based on measuring the received energy level of transmissions from other devices. Any device that is transmitting is, by definition, emitting energy on the radio channel and detection of this energy can enable a device to coexist with any device regardless of technology specific aspects of the signal being transmitted, such as the modulation and coding being used. A more complex method for detection of other devices is the reception of specific digitally modulated and encoded signals and channels. Such signals can additionally provide information about the transmission being detected, such as its duration. However, they also have the disadvantage that they have to be known to neighboring devices, and they are detectable only for the part of the transmission where the signals of a known format are present.
Wi-Fi devices use both received energy as well as the detection of known modulated and encoded signals as part of their LBT procedures. All Wi-Fi devices transmit a preamble at the start of a transmission composed of known reference signals for synchronization and an indicator of the transmission length. A Wi-Fi device does not transmit if, as part of its LBT procedure, a Wi-Fi preamble from a Wi-Fi device is received at an energy level above a threshold of −82 dBm in a 20 MHz channel or any energy is detected above a threshold of −62 dBm, for example, from a non-Wi-Fi device (all thresholds correspond to a 20 MHz bandwidth). Furthermore, an IEEE 802.11ax device may use a higher threshold than −82 dBm when receiving preambles from other IEEE 802.11ax devices that belong to a different network.
This multi-tier threshold regime used by Wi-Fi – different thresholds for different types of neighboring devices depending on whether they are transmitting Wi-Fi preambles and which types of preambles are being transmitted – naturally creates fundamentally different coexistence behaviors depending on the type of neighboring devices. Furthermore, in highly congested environments, preambles are missed more often, creating varying behavior depending on load and interference even among devices that implement the preamble.
The received energy level at which a device defers to another device also has broader implications. It determines the level of reuse of radio resources, i.e., when two devices can transmit on the same resources in frequency at the same time, and therefore is a fundamental determinant of system performance. The level of −82 dBm (and lower) at which Wi-Fi devices defer to each other can lead to significant spectral inefficiency and loss of performance in some environments which has been recognized in the IEEE 802.11 and by Wi-Fi enterprise equipment vendors as well.
The fact that Wi-Fi uses this multi-threshold regime created a critical crossroads in the discussions. The choice was between the adoption of the multi-tier solution used by Wi-Fi, or the adoption of a single common maximum energy detection threshold for all devices. There was a significant effort by the Wi-Fi industry to make 3GPP technologies (LTE-LAA, NR-U) adopt some version of the Wi-Fi preamble or lower its energy detection threshold to −82 dBm. Considering the long-standing success of Wi-Fi, the Wi-Fi industry's strong interest in seeing the Wi-Fi preamble be adopted by 3GPP was understandable.
However, from a 3GPP perspective, the preamble posed multiple problems. The sample rates and OFDM sub-carrier spacings are not compatible with existing 3GPP technologies, thus raising implementation costs. Furthermore, the specification of a preamble with a specific signal format that is based on much older technological capabilities was not an attractive solution from a 3GPP technology evolution perspective. Therefore, it was amply clear that a single common maximum energy detection threshold was the better solution in terms of robustness, fairness, complexity, scalability, adaptability and forward compatibility for cross-technology coexistence. While a solution such as the Wi-Fi preamble may work within a technology domain, for cross-technology coexistence, the costs far exceed the benefits and innovation becomes severely constrained.
This situation naturally led to many years of difficult discussions in multiple fora across multiple 3GPP releases. When the first release of LTE (Rel-13) with operation in unlicensed spectrum was specified, a maximum energy detection threshold of −72 dBm, in between the two thresholds of −62 dBm and −82 dBm used by Wi-Fi, was chosen as a compromise, with the agreement to align all technologies around this single common maximum threshold in the long term. This regime remains in place today even for NR-U despite heavy pressure to adopt the Wi-Fi preamble. Some may see this perseverance as stubbornness. Others may see it as leadership. However, one can simply see it as a commitment to a more robust and practical coexistence mechanism that better serves the long-term interests of all users of unlicensed spectrum across all current and future technologies including Wi-Fi and 3GPP technologies.
This long road has led us, today, to a single common maximum threshold still being the best option for ensuring robust coexistence in unlicensed spectrum. The adoption of such a threshold at -−72 dBm for all devices that operate at the highest permissible transmit power (devices using a lower transmit power are allowed an equivalent increase in the threshold as long as the threshold does not exceed -62 dBm) has now been agreed for the ETSI BRAN harmonized standard being developed for unlicensed spectrum in the 6 GHz band. This regime still allows freedom for any technology to defer to devices at received energy levels lower than −72 dBm, which can yield benefits in many environments, either by the use of technology-specific methods such as a preamble to detect devices of the same technology or by the use of energy detection thresholds lower than −72 dBm to defer to all devices. The recent convergence on this difficult issue will fully realize the potential of robust and fair coexistence between all devices in the 6 GHz band.
The future landscape on unlicensed spectrum
So, what can we expect for the use of unlicensed spectrum in the future?
In terms of activity on the spectrum, the majority of devices will still likely continue to use Wi-Fi. Wi-Fi devices are ubiquitous for everything we do – for home connectivity, for connectivity at meetings, office enterprises – everywhere.
As 3GPP technologies move ahead into 5G and ever-higher spectrum bands, licensed spectrum will remain as a primary focus to serve scenarios and applications where a much higher availability and reliability of service are essential.
Yet, the availability of 3GPP technologies in unlicensed spectrum in the 5 GHz band and the new 6 GHz band which will open up many hundreds of megahertz of spectrum, will provide opportunities to both increase the efficiency of licensed networks by using unlicensed spectrum as a complement, and to serve some scenarios and use cases with 3GPP technologies operating only in unlicensed spectrum in cases where these technologies may be well suited.
Most importantly, the availability of multiple technologies operating in unlicensed spectrum will increase the value that all end users receive from the limited radio spectrum resources of all types available for wireless communications.
There are many in 3GPP, and indeed in the IEEE 802.11 too, who deserve credit for enabling this next chapter in unlicensed spectrum. I would like to express my deep appreciation for the contributions of numerous experts from all the companies that have contributed to the technology and specifications over the last seven years. These contributions do not come without significant investment in the technologies and in 3GPP by the stakeholders involved.
I would also like to express my appreciation for the many dozens of expert researchers, developers and representatives at Ericsson working in concert behind the scenes and on the front lines to secure technological progress. They are the reason why Ericsson has been able to produce such a high amount of technical output in terms of simulations results and analyses and has been prominent in securing access, technological autonomy and future opportunities for 3GPP technologies, including 4G and 5G, in unlicensed spectrum.