Capacity and Power Consumption of Multi-Layer 6G Networks Using the Upper Mid-Band

David L\'opez-P\'erez; Giovanni Geraci; Nicola Piovesan

REVIEW 4 major objections 4 minor 1 cited by

Capacity and Power Consumption of Multi-Layer 6G Networks Using the Upper Mid-Band

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Deploying 6G base stations at traffic hotspots rather than co-locating them with 5G triples median user throughput in the FR3 upper mid-band, to about 301 Mbps with 95th-percentile rates above 1.36 Gbps, at a 33% power-consumption increase.

desk verdict New FR3 multi-layer capacity result is solid and worth refereeing; the 33% power claim is confounded by a baseline swap and needs a like-for-like comparison. read the letter →

arxiv 2411.09660 v1 pith:ZISOVEXH submitted 2024-11-14 cs.NI eess.SP

David López-Pérez , Nicola Piovesan , Giovanni Geraci This is my paper · ORCID

classification cs.NIeess.SP

keywords 6Gnetworksuppermid-bandFR3spectrummulti-layerdeploymentsystem-levelsimulationpowerconsumptioncellreselectionmassiveMIMO

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper is trying to establish that the most effective way to use the upper mid-band (FR3) in 6G is to deploy new low-power base stations at traffic hotspots rather than sharing sites with existing 5G cells, and to steer users onto that layer with priority-based cell reselection. In its system-level simulation of a 4G macro, 5G micro, and 6G multi-layer network, the non-co-located hotspot strategy yields a median user rate of 301.06 Mbps and a 95th-percentile rate of 1.36 Gbps, while co-locating 6G with 5G micro sites reaches 99.64 Mbps. The same model estimates that adding the 6G layer raises network power consumption by 33% over the co-located 4G/5G baseline, with the small pico cells' lower static power keeping the extra cost of new hotspot sites modest. If these numbers hold, operators can choose where to put FR3 capacity before committing to hardware.

What carries the argument

The central object is a multi-layer network model that couples four mechanisms: a standardized urban-macro/urban-micro channel and deployment model with inhomogeneous user density and 19 localized hotspots; beam codebooks for synchronization and channel-state-information reference signals built with two-dimensional DFT precoding (16 SSB beams and 128 CSI-RS beams per 6G cell); RSRP-based initial association followed by priority-based cell reselection, with 6G given top priority and a -108 dBm minimum threshold; and a per-PRB SINR calculation mapped through a mutual-information link model to user rates, plus an analytical base-station power model with transceiver, amplifier, and transmit-power terms. The reselection rule is what moves users onto the 6G layer, and the power model is what converts the deployment choice into the 33% energy cost.

What would settle it

Measure the idle and load power of a real 128-transceiver, 200 MHz, 10 GHz 6G radio and compare it with the model's transceiver, amplifier, and transmit-power terms, which would settle the 33% power claim; a field trial comparing co-located 6G micro cells and non-co-located 6G pico cells at identical hotspots would settle the 301.06 Mbps versus 99.64 Mbps rate gap.

Watch

Extended reading notes

Core claim

On its own terms, the paper claims that FR3 capacity is best unlocked by a non-co-located 6G hotspot layer. Under the modeled traffic load, 6G pico cells placed at the center of 19 hotspots achieve a median downlink user rate of 301.06 Mbps and a 95th-percentile rate of 1.36 Gbps, about three times the 99.64 Mbps median of a 6G micro layer co-located with 5G micro sites, even though the pico cells transmit at lower power. The paper also claims that priority-based cell reselection is as important as placement: without it, users camp on the stronger lower-frequency 4G and 5G cells and the 6G spectrum stays underused. On power, the model puts the 6G premium at 33% over the co-located 4G/5G UMa baseline and attributes the attractiveness of the hotspot strategy to the lower static power of pico radios. The load-bearing method is a multi-layer system-level simulation that combines standard UMa/UMi channel and deployment models, two-dimensional DFT beam codebooks, RSRP-based association, mutual-information SINR mapping, and an analytical base-station power model.

Load-bearing premise

The load-bearing premise is that a power model fitted to 5G base-station data applies to 6G radios with 128 transceivers and 200 MHz carriers without recalibration, so the 33% power figure and the resulting trade-off stand or fall with that extrapolation.

Editorial extensions

If this is right

Operators planning FR3 6G should place new pico sites at traffic hotspots instead of overlaying 6G on existing 5G micro sites; the simulated gain is roughly a tripling of median user rate.
Priority-based cell reselection with 6G as the highest-priority layer must be enabled, or much of the FR3 capacity remains unused despite the new sites.
Adding a 6G layer carries a simulated network power increase of about 33% over the 4G/5G co-located baseline, and using pico radios for new sites partially offsets that cost through lower static power.
Beam-codebook configuration is a first-order lever: with 16 SSB beams and 128 CSI-RS beams per 6G cell, underutilized vertical beams can leave part of the FR3 capacity unrealized.
The multi-layer modeling recipe extends single-layer cellular evaluation to heterogeneous 4G/5G/6G networks, giving a template for testing other FR3 deployment and reselection strategies.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Not stated in the paper: if real 6G radio hardware draws different static power than the 5G-fitted model predicts, the 33% premium could move either way, so the trade-off should be rechecked with physical 6G radio measurements.
Not stated in the paper: the reselection thresholds are uniform per layer, and the paper leaves per-cell optimization to future work; tuning thresholds cell by cell could widen the gap between the hotspot and co-located scenarios.
Not stated in the paper: the full-buffer, round-robin assumptions are favorable to the hotspot comparison, so a trace-driven or bursty-traffic evaluation would test whether the 301 Mbps median survives more realistic loads.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The genuinely new thing here is the multi-layer 4G/5G/6G FR3 system-level comparison. The paper shows that non-co-located 6G hotspot deployments plus priority-based reselection give a median UE rate of 301 Mbps and a 95th-percentile rate of 1.36 Gbps, while co-locating 6G with 5G micro sites gets only 99.64 Mbps. That contrast is not in the cited prior work, and it is a concrete, useful result for the 6G subfield. I believe the capacity result: the SINR and rate equations are standard, the simulator is calibrated to 3GPP models, and the reselection mechanism is clearly described. The paper is also honest about leaving threshold optimization to future work.

What is less solid is the power analysis. The stress-test note is right: the 33% power increase compares the co-located 6G scenario (4G UMa + [5G UMi + 6G UMi]) against [4G UMa + 5G UMa], which differs in two ways at once. The 5G layer moves from 49 dBm macro radios to 44 dBm micro radios, and the 6G micro layer is added. The paper itself says smaller 5G micro radios reduce power, so the 33% number mixes a 5G-layer substitution with the 6G addition. A clean baseline for the co-located 6G case would be 4G UMa + 5G UMi (scenario 5), which is already in the paper. Without that comparison, the abstract's claim that "6G implementation increases power consumption by 33%" is not established. This is fixable, but it is a load-bearing flaw in the power conclusion.

Also, the power model parameters (DTRX, DPA, P0, eta) are not reported, so the decomposition is not independently checkable. The paper reports single simulation runs and no error bars, which is common in this area but worth flagging. The power model is self-cited from one of the authors, but it is used as an input rather than fitted to force the capacity result, so I do not see circularity.

Overall: the capacity finding is likely correct and valuable; the power figure needs re-baselining and parameter disclosure before it can be quoted. This paper deserves a serious referee. I would send it to peer review with a request to fix the power comparison and report the model parameters.

Referee Report

4 major / 4 minor

Summary. The paper presents a system-level simulation model for multi-layer 4G/5G/6G networks operating in FR3 (10 GHz), with seven deployment scenarios that vary co-location, radio type, and site placement. The main reported findings are that a non-co-located 6G hotspot deployment (4G UMa + 5G UMi + 6G HS) yields a median downlink UE rate of 301.06 Mbps and a 95th-percentile rate of 1.36 Gbps, roughly three times the co-located 6G deployment (4G UMa + [5G UMi + 6G UMi], median 99.64 Mbps); that priority-based cell reselection is essential to achieving these gains; and that adding 6G increases network power consumption by 33% relative to a 4G/5G baseline. The paper concludes that strategic non-co-located 6G placement offers the best performance-power trade-off.

Significance. The question addressed is timely and relevant for operators planning FR3 deployments. The paper's strengths are its use of standard 3GPP channel and SINR models, the explicit definition of the rate and power equations, and the systematic comparison of seven deployment configurations, including both co-located and non-co-located 6G options. If the quantitative claims withstand scrutiny, the conclusion that hotspot-targeted 6G pico cells with priority-based reselection are preferable to co-located 6G micro cells would be a valuable, actionable design insight. However, the paper's central power-consumption attribution is currently confounded by a non-like-for-like baseline comparison, the power-model constants are not reported, and the capacity results lack any uncertainty quantification; these issues must be resolved before the headline numbers can be accepted.

major comments (4)

[Section IV.B, Fig. 6; Abstract; Section V] The abstract and conclusion state that '6G implementation increases power consumption by 33%', but the comparison in Fig. 6 is between scenario 4G UMa + [5G UMi + 6G UMi] and scenario [4G UMa + 5G UMa]. These deployments differ in three simultaneous ways: the 5G layer changes from UMa macro radios (49 dBm, 64 TRX) to UMi micro radios (44 dBm, 64 TRX), the 4G layer changes from a multiband 4G/5G macro radio to a standalone 4G macro radio, and the 6G UMi layer is added. The paper itself notes that using smaller 5G micro radios reduces power consumption, so the 33% figure conflates the 5G macro-to-micro substitution and the 4G radio change with the cost of adding 6G. The appropriate like-for-like baseline for adding co-located 6G is scenario 4G UMa + 5G UMi, which the paper already defines; the authors should report that comparison and rephrase the abstract and conclusion accordingly, or remove the 33% claim.
[Section III.C, Eq. (10)] The power consumption model in Eq. (10) drives the headline 33% figure and the power-versus-performance trade-off, but the paper does not report the numerical values of any of the model constants (PBBU, P0, PBB, DTRX, DPA, eta, or the active-TRX count M_PA_ac for each radio type). Without these values the power results cannot be checked or reproduced. In addition, the model of [20] was fitted to 5G base-station data, and applying it to 6G radios with 128 TRXs and 200 MHz carriers is an extrapolation that should be justified, at minimum by stating the parameter values and providing a sensitivity analysis over the extrapolated quantities.
[Section IV.A, Figs. 3-5] The capacity results are reported as point estimates without any indication of the number of independent simulation drops or the variability across realizations of UE locations, shadow fading, and hotspot placement. Quantitative comparisons such as the 15.47% and 23.29% improvements over 5G UMa, and the 301.06 versus 99.64 Mbps median rates, therefore carry no confidence information. The paper should state the number of Monte Carlo drops and report confidence intervals or a variance measure; otherwise it is not possible to assess whether the deployment rankings are statistically significant.
[Section III.A, Radio Units; Table I] There is an inconsistency in the 6G bandwidth: the radio unit list in Section III.A states '6G-only micro radios: 400 MHz, 128 TRXs' and '6G-only pico radios: 400 MHz, 128 TRXs', whereas Table I lists 200 MHz for both the 6G micro and 6G pico radios. Since the achievable rate in Eq. (9) scales linearly with bandwidth, all reported 6G rates depend on which value was actually used in the simulator; please correct the inconsistency and state the exact bandwidth used.

minor comments (4)

[Section IV.A] The claim that co-locating 6G with 5G 'quadruples the capacity of 5G cells' is loose; the bandwidth doubles from 100 to 200 MHz and the CSI-RS beam count doubles from 64 to 128, but the SINR distribution also changes, so the factor four is not self-evident from the text.
[Section III.B, Eq. (2)] In Eq. (2), the transmit power is denoted pssb_{s,b} but should presumably be pssb_{s,c}; please fix the subscript.
[Section III.B] The reselection thresholds of -110 dBm for 5G and -108 dBm for 6G are introduced without justification; a sensitivity analysis over these thresholds would strengthen the claim that priority-based reselection is 'critical', since the magnitude of the gain likely depends on them.
[References] Reference [1] contains a typo ('Recomendation' should be 'Recommendation').

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the capacity and power figures are forward simulation outputs, and the one self-citation is an external power model used as an input rather than a fitted constant or imported uniqueness theorem.

full rationale

The paper's central capacity claims (median 301.06 Mbps, 95th-percentile 1.36 Gbps) are outputs of a forward system-level simulation using 3GPP channel models, explicit radio configurations, and the SINR/rate computations in Eqs. (8) and (9). No parameter is fitted to these target rates, and no equation defines the rates in terms of the claimed result. The power consumption results use Eq. (10) from [20], a prior published model by one of the authors; this is a self-citation, but it is used as an external modeling input and not as a justificatory uniqueness theorem or as an ansatz that already encodes the 33% conclusion. Extrapolating the [20] model to 6G radios without recalibration is a legitimate correctness/assumption concern, but it is not circularity. The '33% increase' comparison also changes the 5G layer from UMa macro to UMi micro while adding 6G, which confounds the attribution of the power increase; that is an apples-to-oranges benchmarking issue, not a derivation that reduces to its own inputs. Overall, the derivation chain is self-contained in the sense that all headline numbers come from explicit simulations rather than from fitting or from self-referential definitions.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The paper's central capacity result relies on the standard 3GPP channel and rate models plus a set of scenario assumptions; the power result additionally depends on parameters inherited from a 5G-fitted model. No new entities are introduced.

free parameters (2)

Power model constants (PBBU, P0, PBB, DTRX, DPA, eta) = Not reported in this paper (from [20])
The 33% power increase result is computed using these constants, which were fitted to 5G base station measurements in [20] and are assumed to hold for 6G radios with 128 TRXs.
Cell reselection thresholds for 5G and 6G = -110 dBm (5G), -108 dBm (6G)
Chosen ad hoc in Section III-B; the paper states optimization per cell is left for future work, but the capacity results depend on these values.

assumptions (3)

domain assumption 3GPP 38.901 UMa/UMi channel model accurately represents FR3 propagation at 10 GHz
Used throughout Section III-A; the paper acknowledges path loss and penetration loss differences but adopts the 3GPP model without FR3-specific calibration.
ad hoc to paper Power consumption model [20] applies to multi-layer, multi-band 6G radios
Equation (10) uses constants from [20] without presenting values or validating for 6G hardware.
domain assumption Round-robin scheduling with full-buffer traffic is representative of network operation
Assumed in equation (9); real schedulers and traffic patterns may produce different rate distributions.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Capacity and Power Consumption of Multi-Layer 6G Networks Using the Upper Mid-Band." pith.science (2026). https://pith.science/paper/ZISOVEXH

@misc{pith2026241109660,
  author       = {Pith},
  title        = {Pith review of: Capacity and Power Consumption of Multi-Layer 6G Networks Using the Upper Mid-Band},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZISOVEXH}},
  note         = {Machine review of arXiv:2411.09660}
}

read the original abstract

This paper presents a new system model to evaluate the capacity and power consumption of multi-layer 6G networks utilising the upper mid-band (FR3). The model captures heterogeneous 4G, 5G, and 6G deployments, analyzing their performance under different deployment strategies. Our results show that strategic 6G deployments, non-co-located with existing 5G sites, significantly enhance throughput, with median and peak user rates of 300 Mbps and exceeding 1 Gbps, respectively. We also emphasize the importance of priority-based cell reselection and beam configuration to fully leverage 6G capabilities. While 6G implementation increases power consumption by 33%, non-colocated deployments strike a balance between performance and power consumption.

Figures

Figures reproduced from arXiv: 2411.09660 by the authors.

**Figure 2.** Horizontal (left) and vertical (right) diagrams of the 64 beams [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Downlink UE rates (no 6G deployments included). [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 5.** Figure 5: Downlink UE rates with and without cell reselection, where [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 6.** Figure 6: Power consumption under different multi-layer deployments. [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Sharing is Caring: Analysis of Hybrid Network Sharing Strategies for Energy Efficient Multi-Operator Cellular Systems
cs.NI 2025-08 conditional novelty 6.0 of 10

Using stochastic geometry and Paris operator traffic, the paper estimates that full network sharing between two mobile operators can save up to about 35 percent of base station energy while meeting QoS targets.

Reference graph

Works this paper leans on

20 extracted references · 18 canonical work pages · cited by 1 Pith paper

[20]

Machine learning and analytical power consumption models for 5G base stations,

N. Piovesan et al., “Machine learning and analytical power consumption models for 5G base stations,” IEEE Commun. Mag. , vol. 60, no. 10, October 2022

work page 2022
[1]

Framework and overall objectives of the future development of IMT for 2030 and beyond,

ITU Recomendation M.2160, “Framework and overall objectives of the future development of IMT for 2030 and beyond,” Nov. 2023

work page 2023
[2]

6G standardization – An overview of timeline and high-level technology principles,

Ericsson, “6G standardization – An overview of timeline and high-level technology principles,” Mar. 2024

work page 2024
[3]

Dahlman et al

E. Dahlman et al. , 5G/5G-Advanced: The next generation wireless access technology. Academic Press, 2023

work page 2023
[4]

What will Wi-Fi 8 be? A primer on IEEE 802.11bn Ultra High Reliability,

L. Galati-Giordano et al. , “What will Wi-Fi 8 be? A primer on IEEE 802.11bn Ultra High Reliability,” IEEE Commun. Mag , vol. 62, no. 8, pp. 126–132, 2024

work page 2024
[5]

Reviewing wireless broadband technologies in the peak smartphone era: 6G versus Wi-Fi 7 and 8,

E. Oughton et al. , “Reviewing wireless broadband technologies in the peak smartphone era: 6G versus Wi-Fi 7 and 8,” Telecommunications Policy, vol. 48, no. 6, p. 102766, 2024

work page 2024
[6]

Artificial neural networks-based machine learning for wireless networks: A tutorial,

M. Chen et al. , “Artificial neural networks-based machine learning for wireless networks: A tutorial,” IEEE Commun. Surveys & Tutorials , vol. 21, no. 4, pp. 3039–3071, 2019

work page 2019
[7]

A survey on fundamental limits of integrated sensing and communication,

A. Liu et al., “A survey on fundamental limits of integrated sensing and communication,” IEEE Commun. Surveys & Tutorials , vol. 24, no. 2, pp. 994–1034, 2022

work page 2022

Show all 20 references

[8]

Cellular wireless networks in the upper mid-band,

S. Kang et al., “Cellular wireless networks in the upper mid-band,” IEEE Open J. ComSoc , vol. 5, pp. 2058–2075, 2024

2024
[9]

7-24 GHz frequency range (Release 16),

3GPP Technical Report 38.820, “7-24 GHz frequency range (Release 16),” Jul. 2021

2021
[10]

6G wireless communications in 7-24 GHz band: Opportunities, techniques, and challenges,

Z. Cui et al. , “6G wireless communications in 7-24 GHz band: Opportunities, techniques, and challenges,” 2024. [Online]. Available: https://arxiv.org/abs/2310.06425

2024 arXiv
[11]

Terrestrial-satellite spectrum sharing in the upper mid- band with interference nulling,

S. Kang et al., “Terrestrial-satellite spectrum sharing in the upper mid- band with interference nulling,” in Proc. IEEE ICC , 2024, pp. 1–6

2024
[12]

Extreme massive MIMO for macro cell capacity boost in 5G- Advanced and 6G,

Nokia, “Extreme massive MIMO for macro cell capacity boost in 5G- Advanced and 6G,” 2023

2023
[13]

Centimeter-wave-MIMO: A key enabler for 6G,

Huawei, “Centimeter-wave-MIMO: A key enabler for 6G,” 2024

2024
[14]

Enabling 6G performance in the upper mid-band through gigantic MIMO,

E. Bjornson et al. , “Enabling 6G performance in the upper mid-band through gigantic MIMO,” arXiv:2407.05630, 2024

2024 arXiv
[15]

Green 5G: Building a Sustainable World,

Huawei Technologies Co., Ltd., “Green 5G: Building a Sustainable World,” Tech. Rep., Aug. 2020. [Online]. Available: https://www. huawei.com/en/public-policy/green-5g-building-a-sustainable-world

2020
[16]

5G energy efficiencies: Green is the new black,

GSMA, “5G energy efficiencies: Green is the new black,” Tech. Rep., Nov. 2020, Accessed on 19/08/2022. [Online]. Available: https://data.gsmaintelligence.com/api-web/v2/ research-file-download?id=54165956&file=241120-5G-energy.pdf

2020
[17]

Study on channel model for frequencies from 0.5 to 100 GHz (Release 18),

3GPP Technical Report 38.901, “Study on channel model for frequencies from 0.5 to 100 GHz (Release 18),” Apr. 2024

2024
[18]

Physical layer procedure for data (Release 18),

3GPP Technical Specification 38.214, “Physical layer procedure for data (Release 18),” Sep. 2024

2024
[19]

Link performance models for system level simulations of broadband radio access systems,

K. Brueninghaus et al. , “Link performance models for system level simulations of broadband radio access systems,” in Proc. IEEE PIMRC, 2005, pp. 2306–2311

2005

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [20]

Machine learning and analytical power consumption models for 5G base stations,

N. Piovesan et al., “Machine learning and analytical power consumption models for 5G base stations,” IEEE Commun. Mag. , vol. 60, no. 10, October 2022

work page 2022

[2] [1]

Framework and overall objectives of the future development of IMT for 2030 and beyond,

ITU Recomendation M.2160, “Framework and overall objectives of the future development of IMT for 2030 and beyond,” Nov. 2023

work page 2023

[3] [2]

6G standardization – An overview of timeline and high-level technology principles,

Ericsson, “6G standardization – An overview of timeline and high-level technology principles,” Mar. 2024

work page 2024

[4] [3]

Dahlman et al

E. Dahlman et al. , 5G/5G-Advanced: The next generation wireless access technology. Academic Press, 2023

work page 2023

[5] [4]

What will Wi-Fi 8 be? A primer on IEEE 802.11bn Ultra High Reliability,

L. Galati-Giordano et al. , “What will Wi-Fi 8 be? A primer on IEEE 802.11bn Ultra High Reliability,” IEEE Commun. Mag , vol. 62, no. 8, pp. 126–132, 2024

work page 2024

[6] [5]

Reviewing wireless broadband technologies in the peak smartphone era: 6G versus Wi-Fi 7 and 8,

E. Oughton et al. , “Reviewing wireless broadband technologies in the peak smartphone era: 6G versus Wi-Fi 7 and 8,” Telecommunications Policy, vol. 48, no. 6, p. 102766, 2024

work page 2024

[7] [6]

Artificial neural networks-based machine learning for wireless networks: A tutorial,

M. Chen et al. , “Artificial neural networks-based machine learning for wireless networks: A tutorial,” IEEE Commun. Surveys & Tutorials , vol. 21, no. 4, pp. 3039–3071, 2019

work page 2019

[8] [7]

A survey on fundamental limits of integrated sensing and communication,

A. Liu et al., “A survey on fundamental limits of integrated sensing and communication,” IEEE Commun. Surveys & Tutorials , vol. 24, no. 2, pp. 994–1034, 2022

work page 2022