5G Dynamic Spectrum Sharing
- Summary
-
Discussion
- How can cellular mobile operators benefit from DSS?
- Instead of DSS, can't we do spectrum re-farming?
- In what bands can we use DSS?
- Is DSS applicable for both FDD and TDD systems?
- What techniques are used in DSS so that LTE and 5G can coexist?
- What IEs are relevant to DSS?
- What's a suitable scheduling algorithm for DSS?
- What are some challenges with DSS?
- Milestones
- References
- Further Reading
- Article Stats
- Cite As
Dynamic Spectrum Sharing (DSS) is a feature introduced in 5G so that both LTE/4G and 5G systems can share the same spectrum. It's dynamic because the allocation between LTE and 5G can be done with a granularity of 1ms. This corresponds to LTE subframe duration.
Spectrum is time-shared. Network takes the decisions. Mobile terminals needn't do any special sensing for DSS. An essential aspect of DSS is backward compatibility. DSS is designed not to affect legacy LTE devices.
Spectrum sharing is actually a more general term in the telecom world. It encompasses many different approaches including Licensed Shared Access (LSA), TV White Space (TVWS), Listen Before Talk (LBT), Automated Frequency Coordination (AFC), and Citizens Broadband Radio Service (CBRS). This article deals with only spectrum sharing between LTE/4G and 5G systems.
Discussion
-
How can cellular mobile operators benefit from DSS? Radio spectrum is scarce. Licensed spectrum is expensive. Legacy LTE operators may not have 5G spectrum licenses. DSS enables these operators to offer their subscribers 5G services over spectrum licensed for LTE. Even for operators with licensed 5G spectrum, DSS provides a more efficient evolution path from LTE to 5G.
In some countries (such as India, Canada and Poland), COVID-19 delayed 5G spectrum auctions. DSS gave operators an option to roll out 5G on existing LTE spectrum.
Legacy LTE devices are not impacted. In the RAN, DSS is only a software upgrade. Operators can reuse existing LTE antennas and radio units for 5G NR as well. Hence operators need not upgrade their RAN hardware.
DSS offers operators an upgrade path from 5G NSA to 5G SA. Initial deployments will involve both LTE and 5G radio accesses via NSA and Dual Connectivity (DC). As LTE is phased out, entire spectrum is allocated to 5G SA deployment.
-
Instead of DSS, can't we do spectrum re-farming? To address spectrum scarcity, regulators sometimes clear bands used by old radio technologies. Those bands are then allocated to newer technologies. This is called spectrum re-farming. An example of this is UHF TV spectrum cleared for 4G services.
Spectrum re-farming is generally a slow process that can take many years. It's also an inefficient use of capacity when a band is re-farmed but enough subscribers haven't migrated to the newer technology. The figure shows how DSS optimizes the allocation. DSS is sometimes referred to as "soft re-farming".
Initial 5G deployments were in unpaired TDD mid-band spectrum such as 3.5 GHz. But much of the spectrum below 6 GHz is paired FDD used by LTE. Operators have to either re-farm this spectrum for 5G or license new spectrum. DSS offers an easier and cheaper route.
Spectrum re-farming has the advantage of reduced inter-cell interference. With DSS, neighbouring LTE cell's CRS can cause interference to serving cell's NR data. Peak data rates with DSS is also reduced due to overheads.
-
In what bands can we use DSS? New 5G spectrum are mostly in mid-bands (3-5 GHz) and high-bands (24-40 GHz). However, better coverage can be achieved in the low-bands (below 3 GHz), which unfortunately are mostly occupied by LTE. Thus, DSS can be applied on these low-bands. We benefit from wide coverage. When combined with Carrier Aggregation (CA), UEs can achieve higher peak rates and enable NSA-to-SA migration. With 1ms scheduling, ultra-low latency applications can be served.
DSS can be used with Dual Connectivity (DC). For example, 5G NR is deployed in a sub-GHz LTE band and the 1-3 GHz band is the LTE anchor. Alternatively, 5G NR in 3.5 GHz TDD band might have uplink coverage issues. DC plus DSS with low-band LTE FDD improves coverage. This combination is called NR-NR DC.
DSS is better suited in some bands such as n5 and n66. It targets LTE bands in the range 700-2700 MHz. Since there's a capacity drop, wide LTE channels such as 10 or 20 MHz are recommended.
-
Is DSS applicable for both FDD and TDD systems? DSS was specified initially for FDD only. This made sense because LTE spectrum were often FDD. DSS in TDD is possible but not really in demand. Many 5G deployments below 6 GHz are already using TDD. For TD-LTE, sub-6 GHz is underutilized or not available. However, MBSFN for DSS purpose can be configured for both LTE FDD and TD-LTE.
Some operators have shown interest in DSS in mid-band TDD. TDD bands above 3 GHz and n41 band don't support 100 kHz raster size, the LTE raster size. Instead 15 and 30 kHz are supported. Hence it's not easy to align NR and LTE carriers.
-
What techniques are used in DSS so that LTE and 5G can coexist? LTE has the capability to configure Multimedia Broadcast Single Frequency Network (MBSFN) subframes. For DSS, six out of ten subframes (1-3, 6-8) can be configured for MBSFN in LTE SIB2. LTE terminals will ignore these subframes. In these subframes, 5G NR Synchronization Signal Blocks (SSBs) are then transmitted.
The figure shows an example of 3 MBSFN subframes, reducing LTE capacity by 7.5%. These subframes don't provide enough capacity. Capacity comes from non-MBSFN subframes that'll be allocated for 5G NR. Scheduler will puncture or shift resource blocks depending on UE capability. Thus, 5G transmissions won't interfere with LTE Cell-Specific Reference Signal (CRS). Moreover, 5G NR's additional PDSCH DMRS is moved from symbol #11 to #12 to avoid conflict with LTE CRS. An MBSFN subframe can contain NR SSB, DMRS and PDSCH transmissions.
Without using MBSFN subframes, it's impossible to accommodate NR SSBs at 15 kHz SCS. This is because there are only 3 contiguous OFDM symbols without LTE CRS but NR SSB needs 4 OFDM symbols. However, at 30 kHz SCS, SSBs can be sent without interfering with LTE CRS.
-
What IEs are relevant to DSS? Rate matching is done with the IE
rateMatchPatternToAddModList
to avoid conflicts with LTE PSS/SSS and PBCH. Likewise, IERateMatchPatternLTE-CRS
informs UE the LTE CRS configuration so that NR PDSCH can be properly received. MBSFN configuration is specified in IElte-CRS-ToMatchAround
in SIB1 or RRCReconfiguration. UE indicates it's capability for DSS in each band with the IErateMatchingLTE-CRS
IE
monitoringSymbolsWithinSlot
is used to configure the position of NR PDCCH so that it doesn't overlap with LTE CRS transmission.In SIB1, IE
frequencyShift7p5khz
informs UE that a 7.5 kHz frequency shift is needed to mitigate inter-carrier interference. In LTE there's no subcarrier at DC and hence LTE does such a shift. This IE ensures carriers in both LTE and 5G NR are aligned.In SIB1 or RRCReconfiguration, IE
n-TimingAdvanceOffset
contains timing advance for DSS.When PDSCH transmission is 13 or 14 OFDM symbols in 15 kHz SCS, additional DMRS is in symbol #11. This conflicts with LTE CRS. IE
additionalDMRS-DL-Alt
is part of UE capability. If supported, network shifts the additional DMRS to symbol #12. -
What's a suitable scheduling algorithm for DSS? 3GPP doesn't specify how DSS scheduling ought to be done. This is the differentiation that infrastructure providers can bring. In fact, each vendor gives their implementation a unique name. In general, the scheduler needs to monitor cell load, number of terminals, the ratio of LTE versus 5G terminals, and types of services being requested.
Scheduling could be done so that both LTE and 5G NR transmissions (date or control) happen in the same subframe. Decisions can be taken at the granularity of 1 ms. In practice, the decision window could be 1-100 ms.
-
What are some challenges with DSS? DSS implementation shouldn't interfere with incumbent LTE transmissions. In particular, 5G NR signals shouldn't interfere with LTE "always on" signals (CRS, PSS, SSS and PBCH). Therefore testers and operators should ensure proper LTE-NR coexistence. Tests should include correct synchronization and carrier performance for both NSA and SA deployments. At the UE, tests should validate correct decoding given the rate matching patterns.
T-Mobile observed that DSS reduces the net capacity of the shared spectrum. However, we should note that T-Mobile has 175 MHz in the 2.5 GHz spectrum range, making it less dependent on DSS for nationwide 5G coverage. Capacity loss depends on vendor implementation. It's typically 10-20%. A good implementation seeks to minimize this loss.
Rate matching for NR PDSCH is possible only for 15 kHz Sub-Carrier Spacing (SCS). For 30 kHz, frequency domain multiplexing is not possible. Only time domain multiplexing is possible.
Milestones
1999
Mitola observes that spectrum could be better utilized if each device could monitor its radio environment and adapt its configuration for current conditions. This is called Cognitive Radio. It's enabled by Software-Defined Radio (SDR). Moreover, spectrum could be flexibly pooled among various services provided all parties conform to formal radio etiquettes.
2006
2007
Zhao and Sadler note that there are many approaches to dynamically sharing spectrum. They propose a taxonomy. One approach they investigate is called Opportunistic Spectrum Access (OSA), aka spectrum overlay. Licensed spectrum is opened up for secondary users. The latter use the spectrum non-intrusively whenever primary users are not using it. Secondary users can detect opportunities by primary transmitters/receivers detection and narrowband/wideband sensing. In 2019, AI-based approaches are proposed.
2017
3GPP approves the first specifications for 5G, called "early drop" of Release 15. This specifies Non-Standalone (NSA) mode of operation. The "main drop" of Release 15 (June 2018) includes Standalone (SA) mode and DSS. A couple of months after the main drop, a DSS specification bug is discovered that could reduce NR DL throughput by 15%. This is resolved soon after. Within 3GPP, DSS was initially named "NR-LTE co-existence".
2019
2020
2020
3GPP finalizes Release 16 specifications. This release improves the capacity of NR PDSCH. In Rel-15, only 2, 4, 7-symbol duration were possible for PDSCH mapping type B. In Rel-16, 10-symbol PDSCH is added. Up to three LTE CRS rate matching patterns can be configured. This is needed since the NR carrier may use a larger bandwidth compared to 10 or 20 MHz LTE bandwidth.
2022
2024
3GPP is expected to finalize Release 18 specifications. As more subscribers transition to 5G, NR PDCCH becomes a bottleneck. NR PDCCH and LTE PDCCH share the first three OFDM symbols of a slot. NR PDCCH can't overlap with LTE CRS. This release will remove the non-overlap limitation. It will also allow a UE to use multiple LTE rate matching patterns to mitigate inter-cell interference (from LTE to NR cells).
References
- 3GPP. 2022. "Revised WID on Enhancement of NR Dynamic spectrum sharing (DSS)." RP-221622, 3GPP TSG RAN Meeting #96, Budapest, Hungary, June 6-9. Accessed 2024-02-15.
- 3GPP. 2024. "Releases." Portal, 3GPP. Accessed 2024-02-20.
- Anritsu. 2021. "Dynamic Spectrum Sharing Implementation Methods and Necessary Testing Practices." Blog, Test Talk, Anritsu, January 26. Accessed 2024-02-13.
- Barb, G., F. Alexa, and M. Otesteanu. 2021. "Dynamic Spectrum Sharing for Future LTE-NR Networks." Sensors, MDPI, vol. 21, no. 12, article no. 4215, June. Accessed 2024-02-12.
- Dahmen, L. 2020. "Intelligent 5G network: How does Dynamic Spectrum Sharing work?" Deutsche Telekom AG, October 27. Accessed 2024-02-15.
- Dano, Mike. 2019. "Another set of 5G standards was just released, but no one really cares." LightReading, April 5. Accessed 2024-02-14.
- ETSI. 2023a. "TR 103 885: Reconfigurable Radio Systems (RRS); Feasibility study on existing spectrum sharing frameworks for temporary and flexible spectrum access." V1.1.2, June. Accessed 2024-02-12.
- ETSI. 2023b. "ETSI and WinnForum Join for New Whitepaper on Spectrum Sharing for Local Private Networks." Press release, ETSI, June 6. Accessed 2024-02-12.
- Ericsson. 2019. "Ericsson Spectrum Sharing – A better way to build 5G spectrum." Ericsson, March 18. Updated 2023-09-04. Accessed 2024-02-12.
- Frenne, M. 2019. "A new standard for Dynamic Spectrum Sharing." Blog, Ericsson, June 19. Accessed 2024-02-13.
- GSMA. 2021. "Spectrum Sharing: GSMA Public Policy Position." GSMA, June. Accessed 2024-02-12.
- KP. 2020. "5G NR: DSS - Dynamic Spectrum Sharing." Blog, How LTE Stuff Works?, November. Accessed 2024-02-21.
- Lin, X. 2022. "An Overview of 5G Advanced Evolution in 3GPP Release 18." IEEE Communications Standards Magazine, vol. 6, no. 3, pp. 77-83, September. doi: 10.1109/MCOMSTD.0001.2200001. Accessed 2024-02-12.
- MathWorks. 2023. "Dynamic Spectrum Sharing for 5G NR and LTE Coexistence." Documentation, MathWorks, Release 2023b, September. Accessed 2024-02-12.
- Mitola, J. 1999. "Cognitive radio for flexible mobile multimedia communications." IEEE International Workshop on Mobile Multimedia Communications, San Diego, CA, USA, pp. 3-10. doi: 10.1109/MOMUC.1999.819467. Accessed 2024-02-14.
- Nokia. 2024. "Dynamic spectrum sharing could be the 5G solution wireless operators are looking for." Nokia. Accessed 2024-02-12.
- Qin, Z., X. Zhou, L. Zhang, Y. Gao, Y.-C. Liang, and G. Y. Li. 2020. "20 Years of Evolution From Cognitive to Intelligent Communications." IEEE Transactions on Cognitive Communications and Networking, vol. 6, no. 1, pp. 6-20, March. doi: 10.1109/TCCN.2019.2949279. Accessed 2024-02-13.
- Qualcomm. 2019. "Ericsson, Swisscom and Qualcomm make huge step towards nationwide 5G coverage in Switzerland." Press note, Qualcomm, November 11. Accessed 2024-02-14.
- Qualcomm. 2022. "5G NR Release 15." Slides, Qualcomm, December. Accessed 2024-02-13.
- Rahman, I., S. M. Razavi, O. Liberg, C. Hoymann, H. Wiemann, C. Tidestav, P. Schliwa-Bertling, P. Persson, and D. Gerstenberger. 2021. "5G evolution toward 5G advanced: An overview of 3GPP releases 17 and 18." Technical report, Ericsson, October 13. Accessed 2024-02-12.
- RantCell. 2024. "Can Dynamic Spectrum Sharing Enable Smooth Transitioning to 5G NR?" RantCell. Accessed 2024-02-12.
- Rayal, F. 2020. "Dynamic Spectrum Sharing: The Pros and Cons." Blog, March 9. Accessed 2024-02-12.
- Roessler, A. 2020. "Your Questions Answered: Dynamic Spectrum Sharing (DSS)." Rohde & Schwarz, April. Accessed 2024-02-12.
- Rowe, M. 2021. "DSS lets 5G and LTE share spectrum." Analog IC Tips, May 10. Accessed 2024-02-15.
- Samsung. 2021. "Dynamic Spectrum Sharing." White paper, Samsung, January. Accessed 2024-02-12.
- WInnForum. 2023. "Feasibility study on existing spectrum sharing frameworks for temporary and flexible spectrum access." WINNF-TR-2011, V1.0.0, WInnForum, January 31. Accessed 2024-02-12.
- Wu, S. 2012. "Network Oriented Spectrum Sharing System." US patent 8,170,576 B2, May 1. Filed 2006-11-15. Accessed 2024-02-14.
- Zemede, M. 2020. "Understanding Dynamic Spectrum Sharing (DSS)." Keysight Design Software, on YouTube, April 18. Accessed 2024-02-12.
- Zhao, Q. and B. M. Sadler. 2007. "A Survey of Dynamic Spectrum Access." IEEE Signal Processing Magazine, vol. 24, no. 3, pp. 79-89, May. doi: 10.1109/MSP.2007.361604. Accessed 2024-02-13.
Further Reading
- Samsung. 2021. "Dynamic Spectrum Sharing." White paper, Samsung, January. Accessed 2024-02-12.
- MediaTek. 2022."Dynamic Spectrum Sharing (DSS)." White paper, MediaTek, January. Accessed 2024-02-12.
- Roessler, A. 2020. "Your Questions Answered: Dynamic Spectrum Sharing (DSS)." Rohde & Schwarz, April. Accessed 2024-02-12.
- Davis, B. 2020. "5G Dynamic Spectrum Sharing (DSS) Technology and Device Testing." Presentation, Anritsu, September. Accessed 2024-02-12.
- Keysight. 2024. "5G Dynamic Spectrum Sharing Testing." Keysight. Accessed 2024-02-12.
Article Stats
Cite As
See Also
- Dynamic Spectrum Sharing
- Licensed Shared Access
- Unlicensed Mobile Access
- MBSFN
- 5G Spectrum
- Cognitive Radio