5G mmWave vs. Sub-6 Spectral Efficiency - a secret sauce of mmWave based 5G Systems (Part-3)

As of January 2021, there were 4.66 billion active internet users worldwide - 59.5 percent of the global population. Of this total, 92.6 percent (4.32 billion) accessed the internet via mobile devices [1]. Among these 4.32 billion mobile internet users apart from Wi-Fi users, there are a significant amount of users are serviced by the cellular networks, in general, using a service called Mobile Broadband (MBB), in the 5G terminology it is called Enhanced?Mobile Broadband (eMBB).

The question to the cellular operators from these 4.32 billion MBB users is, what 5G technology offers to them better than already existing 4G technology? If the 5G cellular operators try to answer this question with terminologies like Sub-6, CBRS/Private Networks, DSS, mMTC, URLLC, and C-V2X, a real expert will know these operators are not answering the question instead they are marketing technology with the terminologies that user don't understand.

These MBB users don't want any of the above terminologies or others which I have missed to mention, all they want is higher speed/throughput and better Quality of Service, independent of the currently available Mobile Applications.

The answer to the above question is ZERO, the 5G systems deployed using Sub-6 spectrum for MBB service offer nothing better than already existing 4G/LTE technology. Some may argue there are many new, better, and more efficient techniques introduced in 5G compared to 4G, but the fact is all these techniques can be introduced in the 4G technology itself. Then a question may arise did the Sub-6 spectrum-based 5G technology will not offer a higher speed than 4G, how about the 3.5GHz band? Yes, it offers, but to get that speed you no need to spend billions of dollars to deploy a 5G network and push the technology to the users, this can be done well by the 4G technology itself.

At this point you will think, then what will the 5G technology offers for MBB users, is it useless? the answer is No. Yes 5G is useful and it offers higher speed/throughput and better quality of service, but how? using the 5G mmWave technology.

I know you will think this is unbelievable, how the 5G mmWave technology will be the answer? is it not cover only the short distance? is it not easily blocked and provides a poor wireless link? is it not for only LOS services like FWA? Yes, all the doubts are correct, but mmWave spectrum is very cheap and it offers very huge bandwidth, and also we have 5G Beam-based technology, AAS antenna, and Analog beamforming based Radio Unit already in the field is it not encouraging?

Many questions will arise, is it possible for the cellular operators to deploy only mmWave based 5G technology? How to deploy it throughout the country, is not the Sub-6 spectrum required for the coverage layer? Not really, the 4G network is already available throughout the country, using the NSA/ENDC mode mmWave can be deployed, but correct, for coverage layer mmWave/LTE technologies are not ready yet. That is because of one of the decisions made by 3GPP when 3GPP decided to reuse the same 4G OFDM technology for 5G ( specifically FR1), they build the standard such a way it need new network devices in both operators side and user side, instead of extending the LTE standard for 5G FR1, same like LTE eMTC/NB-IoT evolved to 5G mMTC. Actually, the 5G FR1 (Sub-6 Frequency range) and FR2 (mmWave Frequency range), should have been 5G Frequency Range 1 Technology (FR1T) and Frequency Range 2 Technology (FR2T), where FR1T should have been 3GPP LTE standard evolved to 5G.

So what happened because of that? LTE will not be able to support more than 20MHz bandwidth (single component carrier (CC) bandwidth), so we can't deploy LTE with 3.5GHz bandwidth like 80/90/100MHz (Yes, in LTE it is possible to do 5CA with each 20MHz CC (20MHz x 5 = 100MHz), whether it will be a feasible solution in mobile device end is the question). So the operators are forced to deploy 5G FR1 technology for the 3.5GHz band and other bands where more than 20MHz bandwidth is available. But going back to our previous point, the deployed 5G Sub-6 networks will not satisfy the requirements of MBB users for the coming next 10 years. Is not the 100MHz bandwidth and Massive-MIMO/MU-MIMO will not serve the MBB user's requirements? Yes and No. The monthly average usage of mobile data in North America is expected to reach 48GB per smartphone, per month, in 2026 [2]. The average traffic per smartphone in the India region stands second-highest globally and is projected to grow to around 40GB per month in 2026 [2]. In 2020, the global monthly mobile data traffic was 51 Exabytes, this will grow to (4x more) 226 Exabytes in 2026 [2].

But mmWave is for FWA and ultra-dense deployment only? Yes, correct in the near future. But in business terms, it is not a bad use case, given 80% of mobile users are indoor and after the Covid-19 pandemic, there will be a significant number of mobile users will be in their homes, so FWA will play a key role. And also ultra-dense use case is where the very high number of mobile users can be served, the higher the user served higher the revenue. But soon the mmWave technology will also find other use cases.

Yes, I accept the current Rel15 based 5G mmWave technology is not fully ready to address the mobile broadband user's need. And also the current techniques are not mature enough to overcome the mmWave challenges. But that should not be a reason to deploy only 5G Sub-6 based technology and avoid the mmWave technology. The USA operators like Verizon and AT&T understood this very well and invested in mmWave spectrum very earlier even before the 5G standardization started and also invested in mmWave technology ecosystem, it paid off, yes there are very smart mmWave repeaters, mmWave RF Chip and Rel16 based mmWave technology startups are coming up in the USA. And also the mmWave techniques in 3GPP are getting mature with each release.

Why spectral efficiency comparison is important?

The Cell bandwidth or spectral efficiency refers to the information rate that can be transmitted over a given bandwidth in a specific communication system. It is a measure of how efficiently a limited frequency spectrum is utilized by the physical layer protocol, expressed in bits per second per Hertz (bps/Hz). For example, if the cell throughput is 100Mbps and it is achieved over 10MHz, the cell spectral efficiency is 100Mbps / 10MHz = 10 b/s/Hz.

There is a general notion, higher the bandwidth higher the throughput, why should someone worry about spectral efficiency?

Yes, the higher the bandwidth higher the throughput, in mmWave, we have a huge bandwidth, but the spectral efficiency limit the throughput, cell spectral efficiency will be affected by the signal quality of the users in the cell, the position of the users within the cell, and the techniques a user is able to enable during the data transmission (for example higher-order modulation, high-rank MIMO requires certain signal quality in general).

Yes, the numerator of the spectral efficiency (S.E) equation, the cell throughput is an important factor that decides the S.E, for example, we can say 100Mbps achieved over 10MHz is better spectral efficiency than 80Mbps achieved over 10MHz. The spectral efficiency of a communication system can be enhanced by packing more information, bits, in a single transmission. This, however, cannot be done unless the quality of the channel is sufficient.

The cell spectral efficiency is the right metric to compare the efficiency of Sub-6 and mmWave physical layer technology because it measures the efficiency of the system in bits per second per hertz, which will not be impacted by the size of the system bandwidth. Typically this will be measured for different channel conditions, for example, if it is measured using simulation, then we can place different users (cell edge/middle/center) in different positions in the cell and use different 3GPP channel models like AGWN and fading channel models (or) if it's measured in the real field we can place different users (cell edge/middle/center) in different channel conditions like LOS and NLOS.

Another question will arise, the mmWave cells are small cells, whether comparing Sub-6 large cells against the mmWave small cells are valid or is it really make sense to do it. Yes, the comparison is valid, because mmWave is happen to be small cells due to physics reasons, if we want to evaluate how the physical layer technology efficiently use the mmWave spectrum, the comparison between Sub-6 and mmWave technology will help us to understand how good is the mmWave physical layer technology. And to compare the physical layer technology the Spectral efficiency is the right metric.

5G mmWave vs. Sub-6 Technology Spectral Efficiency Comparison

Source#1 [3]

These are the T-Mobile Spectral Efficiency (S.E) numbers given to FCC [3]. Average Spectral Efficiency is used to represent how a Typical network cell will operate under an average network traffic loading. These numbers are obtained from T-Mobile vendors simulation as we have described in the above section.

Conclusion: Above table shows the mmWave S.E is almost double of mid-band S.E, here mMIMO means AAS antenna with large phased array antenna, not like the Sub-6 64T64R systems with 64 RF chains doing digital beamforming. It is not mentioned but most probably 4x4 SU-MIMO in mmWave.

Source#2 [4]

Myth: The 5G mmWave based systems will work in LOS only, What is so great if its spectral efficiency is more than Sub-6 systems? Even if it works in NLOS the throughput will be very less, and the NLOS condition in which it works is also very restrictive, it will not work penetrating across a building or dense foliage or multiple walls.

Yes, the above myth seems to be correct, but we will check the facts. The mmWave systems work in NLOS because of the strong reflected RF signal property, and also it can achieve moderate throughput. Yes, it works in NLOS with restrictions, it will not work as Sub-6, yes the mmWave sees even the objects that the Sub-6 penetrates as blockages. But there is no solution to blockages? that is the challenge in mmWave, this can be solved using AI/ML or Multi-TRP or brilliant beam management techniques or repeaters, but all these solutions are in paper only, except the repeaters. The solutions to the mmWave challenges are evolving, it will get better day by day through 3GPP and proprietary solutions, the market differentiator for mmWave product startups is an open secret, "solve the mmWave known challenges to some acceptable level for cost-effective deployment".

The sample cell throughput numbers from field testing by Ericsson in LOS and NLOS scenarios are given below. The throughput results are similar under both cases (LOS and NLOS) above ~300m but for close to the cell location (~150m), the LOS shows a significant advantage in terms of throughput performance. Cell radius ~33% lower in case of NLOS.

NLOS DL Peak S.E = 3.5 b/s/Hz

NLOS UL Peak SE = 1.25 b/s/Hz

Conclusion: Ericsson peak S.E numbers are a good marker to show mmWave S.E numbers are good for the NLOS channel also.

Source#3 [6]

This GSMA report [6], uses the below technical assumptions to analyze the economics of mmWave, the report assumes the mmWave S.E is better than the 3.5GHz band for three different use cases, Dense Urban, FWA, and Indoor. The below table shows the technical assumption of dense urban use case.

Dense Urban:

Conclusion: Even though the GSMA report is just a technical assumption, the assumption may be based on solid background information. In the above assumptions, the mmWave S.E is significantly more than the 3.5GHz band.

Source#4 [4]

The theoretical peak mmWave DL spectral efficiency based on the chart above comes to ~4 b/Hz assuming 64QAM and single layer. It should also be noted that recently Signals Research Group performed practical mmWave measurements in the field using Verizon’s mmWave network in Minneapolis and Chicago [5]. A total of 4 component carriers, each 100MHz wide were configured. The peak DL spectral efficiency of that mmWave was noted to be ~2b/s/Hz. This shows that the practical case peak spectral efficiency (2b/s/Hz) may be ~50% lower than the assumed theory.

Conclusion: The mmWave practical peak S.E is 2 b/s/Hz (for 2x2 MIMO), which is better or comparable with Sub-6 with 2x2 MIMO and 64QAM.

Conclusion

In this article in the beginning we discussed the role and importance of 5G mmWave technology, and later discussed the spectral efficiency of mmWave vs. Sub-6 to establish a point, not only the bandwidth of mmWave technology is huge, but it also uses the bandwidth/spectrum efficiently, compared to 5G Sub-6 technology. So, the mmWave bandwidth or spectral efficiency is also going to be a secret sauce for mmWave based 5G systems' success.

Reference:

[1] ? Internet users in the world 2021 | Statista

[2] Mobile data traffic forecast – Mobility Report - Ericsson

[3] Oct 11 2018 Ex Parte (Public).pdf (fcc.gov)

[4] IWPC_062019.pdf (skyworksinc.com)

[5] White Paper 012020.indd (qualcomm.com)

[6] The economics of mmWave 5G (gsmaintelligence.com)