Leveraging the unlicensed Band: How MulteFire overcomes the spectrum conundrum
Operators and enterprises foresee an extraordinary increase in data traffic by 2020. Unfortunately, today's scarcity of spectrum is unlikely to support this growing problem. Users have turned to Wi-Fi to help fill the spectrum gap but as vital as the system is, it still has service limitations. What if it could deliver massive outdoor service, with widespread coverage, and mitigated interference? What if it could combat the future data deluge to ensure sufficient throughput? If Wi-Fi alone is incapable of these feats, how could it be augmented to truly deliver stellar service? Users demand (and deserve) more.
Key Research Areas:
MulteFire's key challenge: Making the jump from licensed to unlicensed to augment Wi-Fi.
Qualcomm Research's goal in designing MulteFire was to leverage existing LTE technology within the unlicensed spectrum to deliver enhanced capacity, improved coverage, seamless mobility, and superior quality of service in order to better augment Wi-Fi. We created the system based upon 3GPP standards, specifically, Licensed Assisted Access (LAA) in Release 13 for downlink and enhanced Licensed Assisted Access (eLAA) in Release 14 for uplink. We began development of LAA in 2014 and the leap from LAA to MulteFire the following year involved the removal of dependence on licensed anchor and engineering the system to work completely on an unlicensed band.
Conquering Migration Complications
Creating MulteFire presented significant challenges for engineers. First, we needed to transform the existing LTE technology that was designed to work in a licensed spectrum and migrate it to the unlicensed spectrum. Unlicensed band is not reliable as its channel availability is not always guaranteed. Second, LAA was not designed for standalone operation on unlicensed band and always had a licensed anchor carrier. So operating there was easier because licensed spectrum was always available for critical functions and unlicensed band was just used opportunistically. So, learning to perform critical functions in the unlicensed band was a real test.
We overcame these obstacles through algorithmic research and restructuring the way we transmitted control signals, ranging from control messages facilitating data transfers to recovering from radio link failures. All of these processes were migrated to the unlicensed band. To facilitate this, we introduced the DMTC (DRS Measurement Timing Configuration) window, a new technique that allows MulteFire to transmit but with minimal interference to other unlicensed technology including Wi-Fi. Additionally, the periodicity of discovery signals is very sparse. This allows us to access channels occasionally, transmit discovery and control signals, and then vacate the channels.
Overcoming the Handover Dilemma
Another critical challenge we faced was making handovers work in the unlicensed band. For example, in a licensed band, where a caller may be walking through a shopping mall, the user's call is automatically handed over from base station to base station seamlessly, thus preventing calls from being dropped. This is much more difficult in the unlicensed band. Leveraging the DMTC window, we developed cutting-edge MulteFire algorithms which search and decode reference signals in unlicensed band from neighboring base stations in order to know which base station would best for serving the user. As the caller moves past one base station, their UE sends a measurement report to it, triggering a handover at the right moment, and transferring the caller (and all of their content and information) to the next base station. Qualcomm Research was the first to complete outdoor handovers on an unlicensed band.
The UE and two base stations share information to ensure a seamless handover which triggered as Base Station 1 receives a measurement report from the UE.
Improving upon Wi-Fi communication with Channel Quality Indication (CQI).
MulteFire provides constant Channel Quality Indication (CQI) feedback from the user to the Access Point (AP), allowing it to know the user's precise channel and interference conditions, similar to LTE and LAA/eLAA. For example, this feedback is leveraged by the AP's scheduler to determine which time slot to utilize for communication with a particular UE as well as how much resource (e.g. frequency PRBs, power) and coding, modulation, and MIMO scheme to use.
Engineering MulteFire and Wi-Fi to co-exist via LBT.
Since LTE traditionally operated in licensed spectrum and Wi-Fi operated in unlicensed bands, coexistence with Wi-Fi or other unlicensed technology was not considered when LTE was designed. In moving to the unlicensed world, we modified our LTE waveform and added algorithms in order to perform Listen Before Talk (LBT). This allows us to respect unlicensed incumbents including Wi-Fi by not just acquiring a channel and immediately transmitting. Our prototype implementation supports LBT and the detection and transmission of WCUBS (Wi-Fi Channel Usage Beacon Signal) for ensuring coexistence with Wi-Fi neighbors. We designed MulteFire to “hear” a neighboring Wi-Fi base station's transmission (because it's all unlicensed spectrum). MulteFire listens first, and autonomously makes the decision to transfer when there is no other neighboring Wi-Fi transmitting on the same channel. This technique ensures co-existence between MulteFire and Wi-Fi.
Additionally, we adhere to the unlicensed rules and regulations set by 3GPP and the European Telecommunications Standards Institute (ETSI), which mandates the -72dBm LBT detection threshold. This further helps us de-conflict with Wi-Fi. MulteFire's LBT design is identical to the standards defined in 3GPP for LAA/eLAA and complies with ETSI rules.
Making history: Qualcomm Research demonstrates MulteFire's ability to augment Wi-Fi.
We have successfully demonstrated MulteFire's potential over the past year, proving that the system can improve a network's throughput performance. These events marked a historical milestone for our MulteFire program as Qualcomm Research was the first to demonstrate this capability.
Mobile World Congress, February 2016
During this event, our engineers proved that MulteFire could not only coexist with Wi-Fi, it could actually help Wi-Fi improve its overall throughput. The demo began with two pairs of state-of-the-art Wi-Fi nodes (802.11ac with 2x2 MIMO and LDPC codes), all transmitting and supporting simultaneous DL/UL traffic, operating on a 20 MHz channel in a 5 GHz spectrum and above energy detection (meaning they were close together and could hear each other nicely). After they all achieved a certain network throughput, we replaced the two Wi-Fi nodes with two MulteFire nodes, which soon achieved about 30% more throughput. Additionally, the two Wi-Fi nodes' individual throughputs remained the same or increased slightly and the overall throughput of the network increased by nearly 30%. Thus, the demonstration proved that MulteFire not only can coexist with Wi-Fi, it actually increases its performance.
Sharing the medium with more spectrally efficient neighbors like MulteFire not only improved Wi-Fi’s performance, but proved that MulteFire is a better neighbor to Wi-Fi than Wi-Fi itself.
Qualcomm Research Campus, November 2016
MulteFire made history again in this event, which repeated the February 2016 demonstration, but conducted it outside in a real-world environment. During the demo, the four Wi-Fi nodes achieved 25 Mbps, then were later replaced by MulteFire nodes, which achieved 52 Mbps, delivering more than 2x gain, greatly increasing the network's spectral efficiency.
Mobile World Congress, February 2017
During this leading-edge demonstration, our engineers will prove that MulteFire can coexist fairly with Wi-Fi, while providing improved user experience in terms of throughput, coverage, and seamless mobility. This live over-the-air (OTA) demo will kick off from within Qualcomm Research HQ, with four pairs of Wi-Fi APs and devices sharing the same channel in 5GHz unlicensed spectrum in a mixed deployment. For example, some nodes will detect each other above the Energy Detect (ED) threshold (-62dBm) while others will detect below ED. Next, watch the overall network throughput increase and the OTA packet collisions decrease as we subsequently convert two Wi-Fi pairs into MulteFire. Finally, we’ll convert the remaining two pairs to MulteFire, resulting in significant gains in overall network throughput compared to the Wi-Fi baseline.
MulteFire will then perform its first live, outdoor OTA demo, where we’ll run a head-to-head comparison between Wi-Fi and MulteFire’s coverage. Watch as a base station positioned at our HQ transmits both Wi-Fi and MulteFire to our nearby roving vehicle equipped with Wi-Fi and MulteFire devices. As the vehicle moves away from the base station/AP, you’ll see MulteFire’s range go much further beyond where the Wi-Fi call drops, showing a significant gain in coverage. In our second outdoor OTA demo, we’ll show seamless handovers in a multi-node MulteFire deployment while sharing the channel with other Wi-Fi nodes.
MulteFire Alliance and Qualcomm Research: Setting the standard for MulteFire
In 2015, Qualcomm and its industry partners formed the MulteFire Alliance, an international consortium focused on creating global technical specifications for MulteFire, instituting a product certification program and promoting a global ecosystem for MulteFire. We recently led the way in delivering the Alliance's Release 1.0 specification, which was tightly aligned with 3GPP specification standards. Combining Qualcomm Research's success with the readiness of the Release 1.0 specification sets the stage for MulteFire user trials.
MulteFire: A critical stepping stone to enabling 5G NR Shared Spectrum (NRSS).
Although MulteFire operates in 5 GHz and there are many channels to operate from, the spectrum demand continues to surge so new solutions must be found to deal with all the users. In response to this, the FCC released 150 MHz of spectrum in the U.S. at 3.5 GHz. Also known as the Citizens Broadband Radio Service (CBRS), this spectrum is currently occupied by the U.S. military and its bandwidth remains largely unused, allowing the large mobile carriers access to it for a fee through PAL (Priority Access Licenses) or for free via GAA (General Authorized Access). The carriers leverage this spectrum as additional bandwidth for their customers. If carriers share GAA spectrum with each other, they can access a wider block of spectrum and the OTA transmission time for a given user burst would reduce significantly, enabling better user experiences. Unfortunately, the tradeoff is that the benefits diminish as the system becomes saturated with users, creating more contention for this medium, including the possibility of channel collisions. The performance of OTA sharing in this region can be further improved with new features in 5G NRSS technology.
We have leveraged our advancements in MulteFire in order to conceptualize our design for a 5G NRSS prototype. This system is derived from our sub-6 GHz 5G NR prototype but we are taking its source code and branching it off by adding techniques we pioneered with MulteFire including LBT, channel assessment, and the sending/decoding of preambles. Additionally, it will add new techniques such as high-order MIMO, enabling an even better data rate. Designed to operate in the shared/unlicensed band, it will serve as a testbed to drive 5G standardization in spectrum sharing.
If you find the work we're doing in MulteFire to be exciting, and you have a technical background in MulteFire design, Wi-Fi design, or 5G technologies, we'd love to hear from you. Please visit us at www.qualcomm.com/company/careers to submit your resume'. When creating your Qualcomm profile, please enter the activity code, "MulteFire".