Pith. sign in

REVIEW 2 major objections 7 minor 21 references

Performance Analysis of Traditional and Network Coded Transmission in Infrastructure-less Multi-hop Wireless Networks

T0 review · 2 major / 7 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 Network-coded relays give 50–100% higher capacity than store-and-forward in modeled infrastructure-less multihop networks.

desk verdict A simple simulation study with a textbook claim, undercut by a capacity-formula error that likely inflates the network-coding advantage. read the letter →

arxiv 2411.14539 v1 pith:4G2FHLMR submitted 2024-11-21 cs.NI eess.SP

Muhammad Ali , Alister Burr This is my paper · ORCID
classification cs.NIeess.SP
keywords networkcodinginfrastructure-lesswirelessnetworksmulti-hoprelaystransmissionschedulingcapacityboundSINRXORthroughputcomparison
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 sets out to show that, in an infrastructure-less multi-hop wireless network, relaying data with bitwise-XOR network coding yields a substantially higher capacity bound than traditional store-and-forward relaying under identical interference conditions. Its simulations cover a deliberately simple two-row topology with one or two simultaneous communication streams and two to five hops, with a periodic transmission schedule of period $Z$ controlling which nodes may transmit. Across every tested configuration, network-coded transmission delivered the higher throughput, with gains of 50% to 100% for a single stream and 52% to 88% for two coexisting streams. If the result holds in practice, network coding plus scheduling is a low-complexity way to push more data through interference-limited tactical and disaster-response networks without changing topology or transmit power.

What carries the argument

The load-bearing mechanism is the combination of a periodic schedule of period $Z$ with relay-side bitwise XOR combining. The schedule limits simultaneous transmitters so that in timeslot $i$, nodes $i, i+Z, i+2Z, \dots$ transmit, which controls physical interference. In traditional mode, forward and reverse schedules alternate and the stream delivers one packet in each direction every $2Z$ timeslots; in network-coded mode, a relay XORs a newly received packet with a stored one, so one packet can travel left and one right in the same period, delivering two packets every $Z$ timeslots. The XOR algebra is just the identity $A \oplus B \oplus B = A$, which lets each source recover the other's packet from the combined transmission. This mechanism converts a scheduling choice into a throughput difference: fewer simultaneous transmitters raise SINR, while network coding converts the schedule into a doubled delivery rate per timeslot.

What would settle it

Re-run the same two-row topology as a packet-level experiment: measure the steady-state number of packets delivered per direction per schedule period with and without small XOR headers and finite relay buffers. If network-coded transmission cannot sustain one packet per direction every $Z$ timeslots, or if its measured throughput gain over traditional relaying falls below the model's 50% for a four-hop single-stream network, the zero-overhead steady-state assumption is contradicted.

Watch

Extended reading notes

Core claim

The central discovery is a capacity-doubling scheduling effect. In traditional relayed transmission, forward and reverse traffic alternate under the half-duplex constraint, so two packets, one in each direction, require $2Z$ timeslots and the stream capacity is $(R_f+R_r)/(2Z)$. With network coding, relay nodes transmit a bitwise XOR of a packet from the left and a packet from the right, so both directions advance in the same schedule period and two packets arrive every $Z$ timeslots, giving capacity $(R_f+R_r)/Z$. The model computes each node's SINR from all simultaneous transmitters, converts it to a per-node rate with the capacity formula, and takes the minimum rate along the route in each direction. In all simulations the network-coded schedule produced higher capacity than traditional relaying, and the gap was largest for long routes; the advantage also persisted when a second stream added cross-stream interference, with smaller percentage degradation than traditional relaying.

Load-bearing premise

The network-coding advantage rests on assuming a relay always has a stored packet to XOR with the newly received packet every schedule period, with no protocol overhead; if a relay must wait or send extra headers, the factor-of-two rate gain shrinks.

Editorial extensions

If this is right

  • For a single communication stream with two to five hops, the simulated network-coded capacity exceeds traditional relaying by 50% to 100% at the respective optimum scheduling periods.
  • For two coexisting streams, the network-coded advantage remains 52% to 88%, showing that network coding is less degraded by cross-stream interference than traditional relaying.
  • The optimum scheduling period is topology-dependent: traditional relaying favors $Z=2$ at two hops and moves to $Z=3$ or $Z=4$ on longer routes, while network coding favors $Z=4$ for one stream and $Z=3$ for two streams in the tested range.
  • Increasing the number of hops lowers the attainable per-node data rate under both schemes, yet the relative gain of network coding persists across hop counts.

Reading between the lines

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

  • Editorial inference: the modeled factor-of-two timeslot advantage is an upper bound; real XOR operations, headers, and relay buffering overhead would erode the 50–100% gain, so the numbers should be read as best-case for a clean protocol.
  • Editorial inference: the mechanism is not tied to the two-row layout; any topology where relays can pair left- and right-moving packets should show a similar scheduling advantage, provided SINR is computed from simultaneous transmitters.
  • Editorial inference: a direct extension would be to add a third parallel stream; the paper's observed pattern that network coding degrades less than traditional relaying with added interference suggests the relative advantage could persist, but additional interference rows could also change the optimum schedule period.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 7 minor

Summary. This paper presents a simplified analytical model, implemented in MATLAB, that compares traditional relaying (TR) and network coding (NC) in infrastructure-less multi-hop wireless networks. The network layout consists of one or two parallel rows of equally spaced nodes, and the model computes per-link SINR using an inverse-power path-loss law, maps SINR to capacity via the Shannon formula, and aggregates per-direction rates over a scheduling period to obtain a network capacity figure. The central claim is that NC achieves 50% to 100% higher Shannon capacity than TR in all simulated configurations (hop counts 2 to 5, scheduling periods 2 to 5, one or two streams). The paper also reports that increasing hop count reduces capacity and that two coexisting streams degrade performance relative to a single stream.

Significance. The paper's main strength is that it provides a concrete, parameterized calculation chain from physical-layer interference to end-to-end capacity in a simple multi-hop setting, and it explicitly incorporates transmission scheduling and interference, which many qualitative treatments of network coding omit. If the capacity aggregation and steady-state scheduling assumptions were correct, the quantitative claim of 50–100% improvement would be a useful baseline for tactical ad-hoc network design. However, the load-bearing capacity formula for NC is not correct as written, and the NC scheduling algorithm is not defined or verified for the full parameter sweep. The paper also does not provide code, a complete schedule specification, or packet-level verification, which limits reproducibility. The qualitative direction of the result (NC can outperform TR under interference) is plausible, but the specific quantitative claims in the abstract and Section VI are not established by the current model.

major comments (2)
  1. [III.A] The capacity normalization for NC is not justified. The text states that the NC capacity per timeslot is (R_f + R_r)/Z, whereas the described NC operation has a relay transmit a single XOR-coded packet to both adjacent receivers in the same timeslot, as in Scenario 2 (§III.D, TS4). A single broadcast transmission cannot support two different rates toward the two receivers; the achievable common rate is min(R_f, R_r). The total useful throughput in steady state is therefore at most 2·min(R_f, R_r)/Z, not (R_f + R_r)/Z. The two directions can have substantially different SINR because the interference environment differs between forward and reverse transmissions, as reflected in the different latencies in Eqs. (13) and (14). The reported 50–100% gains in Table 4 are thus not established by the model as written. The authors should either correct the NC capacity formula, or justify that the two directions are scheduled in separate timeslots (which contradicts the described XOR broadcast), or provide a worst-case analysis using min(R_f, R_r).
  2. [III.C and III.D] The steady-state assumption that NC sustains one packet per direction per Z timeslots for all N_o and Z is only illustrated by the hand-run example N_o = 5, Z = 4 in Scenario 2 and Figure 7. No general proof or packet-level simulation is provided for the other configurations used in the results (N_o = 3..6, Z = 2..5). Moreover, the NC schedule is not formally defined in the way TR is defined by Eqs. (7) and (8). The example interleaves reverse-direction transmissions (e.g., node 2 transmitting to node 1 in TS10 of Figure 7) in a manner that is not captured by the 'simple sequential forward schedule' mentioned in §III.C. Without a precise schedule for each timeslot and a proof of the one-packet-per-direction-per-Z-timeslot property, the capacity results for the entire parameter sweep in Table 4 are unsupported. The authors should provide a general scheduling algorithm for NC and verify the steady-state rate for all simulated N_o and Z values, either analytically or by simulation.
minor comments (7)
  1. [General structure] The Introduction promises a Section IV on the process for computing SINR, but the manuscript jumps from Section III.D directly to Section V. The SINR computation is only referred to in passing and via Eq. (6), which hampers reproducibility.
  2. [Figures] Figures 6 and 7 are each used twice: once in Section III.D (scheduling scenarios) and again in Section V (capacity plots). This creates confusion and should be corrected by renumbering.
  3. [Table 4] The values in the last column of Table 4 have inconsistent signs. For example, for N = 3, the TR row shows -15% but the numbers imply (C_OS - C_TS)/C_TS = +18%. The authors should check the sign convention and the associated formulas for this column.
  4. [References] The reference list jumps from [18] to [21]; entries [19] and [20] are missing. In addition, some cited works date from 2005–2015 and more recent literature on network coding and scheduling could be cited.
  5. [Equations] Equation numbering skips: Eq. (2) and Eq. (5) are not present in the text, though Eq. (1), (3), (4), and (6) are referenced. This should be corrected for clarity.
  6. [Section III.D] There is a typo in Scenario 2: 'node 4 transmits the NC packet A⊕B to both node and node 5' should read 'to both node 3 and node 5'.
  7. [Abstract and Section VI] The claim that NC outperforms TR 'under the same network settings' is imprecise, because Table 4 compares each scheme at its own optimal scheduling period Z. The comparison of optima is a legitimate approach, but it should be stated explicitly so that the reader does not assume identical Z.

Circularity Check

1 steps flagged · score 6.0 of 10

NC advantage is largely built into the capacity-per-timeslot definition (Z vs 2Z denominator), so the headline 50–100% gain is partly a restatement of the model's own normalization rather than an independent result.

  1. self definitional [Section III.A, 'Simulator operation'; also Section III.D, 'Scenarios to illustrate scheduling']
    "The sum of the two data rates in each direction, divided by the schedule period Z is the capacity per timeslot of this communication stream for the NC case whereas for the TR case it is divided by 2Z. ... Hence the rate is 2/Z which will be multiplied with the data rate calculated for each communication stream to find the actual throughput."

    The paper's headline result—'IMWNs with NC transmission display a higher Shannon capacity bound (between 50% to 100%) than those with TR transmission'—is a direct consequence of this definitional normalization. For identical per-direction Shannon rates, the NC capacity is (R_f+R_r)/Z and the TR capacity is (R_f+R_r)/(2Z), so the definition alone inserts a factor of two before any SINR computation. The later percentages in Table 4 (50–100%) are computed from these formulas, not independently measured. The SINR-dependent rates can alter the ratio, but the direction and rough magnitude of the claimed advantage are fixed by the chosen denominators, so the prediction reduces to the model's own capacity definition.

full rationale

This paper is self-contained and contains no fitted parameters and no load-bearing self-citations; the SINR computation and Shannon-link calculations are legitimate model steps. However, the central comparison is partially circular. The capacity-per-timeslot definitions in Section III.A divide by Z for NC and by 2Z for TR, which by construction makes NC twice as efficient whenever the per-direction Shannon rates are comparable. The later statement that NC is 50–100% better is therefore largely an output of that normalization rather than an emergent finding from interference physics. The paper does provide a worked scheduling example (N_o=5, Z=4) and computes SINR-dependent rates, so the result is not purely equivalent to a single fitted parameter; but the decisive factor-of-two advantage is assumed in the metric, not derived across the full parameter range. This warrants a partial-circularity score of 6 rather than a higher score, because independent SINR effects still enter the model and can modulate the final percentages.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The model contains no statistical fits and no invented physical entities. All quantitative inputs are user-selected from standard engineering ranges. The central comparison is driven by two ingredients: the SINR calculation based on additive interference, and the schedule-normalization factor (Z versus 2Z) that encodes the assumed bidirectional efficiency of network coding. The domain assumptions, especially zero XOR overhead and perfect synchronization, are load-bearing for the reported gains.

free parameters (4)
  • Path loss exponent eta = 4
    Chosen to model a cluttered environment; higher eta makes distant interferers weaker, directly shaping the NC versus TR gap.
  • Route separation d1 = 300 m
    Selected from the range 300-1000 m; at 300 m with hop length 100 m, inter-stream interference is significant.
  • Scheduling period Z = 2 to 5 (swept)
    Swept to find the optimum; the reported 50-100% NC advantage is the envelope over Z values.
  • Noise figure F = 4 dB
    Standard value chosen from the 2-6 dB range; affects absolute rates but not the relative TR/NC comparison.
assumptions (6)
  • standard math Shannon-Hartley theorem gives the achievable link rate from SINR.
    Used in Section II.C.4 and Section IV to convert SINR values into bit rates.
  • domain assumption Inverse power law path loss (Eq. 1) with eta = 4 models both signal and interference.
    Section II.C.2; the chosen PLE of 4 is a strong environmental assumption, and no small-scale or large-scale fading is modeled.
  • domain assumption Interference powers add linearly, captured by SINR = P_RX / (I_x + P_n).
    Section III.A, Eq. (6); assumes no capture effect, no interference cancellation, and perfectly additive interference.
  • domain assumption All nodes are half-duplex and follow a synchronized, periodic schedule with period Z across every stream.
    Section II.B; the entire interference calculation depends on this global synchronization.
  • domain assumption Network coding XOR operations incur zero protocol overhead and relays always know which packets to combine.
    Section III.C; the latency and throughput formulas assume free XOR combining and perfect header information.
  • domain assumption Nodes are stationary and placed in a rigid two-row grid with uniform spacing.
    Section II.A and II.C.1; distances are fixed, which removes mobility and random topology effects from the comparison.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Performance Analysis of Traditional and Network Coded Transmission in Infrastructure-less Multi-hop Wireless Networks." pith.science (2026). https://pith.science/paper/4G2FHLMR

@misc{pith2026241114539,
  author       = {Pith},
  title        = {Pith review of: Performance Analysis of Traditional and Network Coded Transmission in Infrastructure-less Multi-hop Wireless Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4G2FHLMR}},
  note         = {Machine review of arXiv:2411.14539}
}
read the original abstract

Infrastructure-less Multi-hop Wireless Networks are the backbone for mission critical communications such as in disaster and battlefield scenarios. However, interference signals in the wireless channel cause losses to transmission in wireless networks resulting in a reduced network throughput and making efficient transmission very challenging. Therefore, techniques to overcome interference and increase transmission efficiency have been a hot area of research for decades. In this paper two methods for transmitting data through infrastructure-less multi hop wireless networks, Traditional (TR) and Network Coded (NC) transmission are thoroughly examined for scenarios having one or two communication streams in a network. The study has developed network models in MATLAB for each transmission technique and scenario. The simulation results showed that the NC transmission method yielded a better throughput under the same network settings and physical interference. Furthermore, the impact of increasing numbers of hops between source and destination on the network capacity and the communications latency was also observed and conclusions were drawn.

Figures

Figures reproduced from arXiv: 2411.14539 by the authors.

Figure 1
Figure 1. Wireless Infrastructure less Multi-hop Network [2] The architecture and functioning of IMWNs have been discussed at length in previously published papers that also highlight the problems that arise in such networks [2-4]. When a receiver in a wireless multi-hop network is out of the coverage/transmission range of the transmitter, relaying is needed. Routing facilitates the relaying of data packets through intermedia… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 20 canonical work pages

  1. [1]

    Ubiquitous robust communications for emergency response using multi-operator heterogeneous networks

    Fragkiadakis, A.G., Askoxylakis, I.G., Tragos, E.Z. et al. “Ubiquitous robust communications for emergency response using multi-operator heterogeneous networks”. Journal on Wireless Communications and Networking vol 13 (2011). https://doi.org/10.1186/1687-1499- 2011-13

    work page doi:10.1186/1687-1499- 2011

  2. [2]

    A Cross- Layer Design for a Multihop, Self-Healing, and Self-Forming Tactical Network

    Nosheen, Irum, Khan, Shoab A., Ali, Umar, “A Cross- Layer Design for a Multihop, Self-Healing, and Self-Forming Tactical Network”, Wireless Communications and Mobile Computing, 2019, https://doi.org/10.1155/2019/1523906

    work page doi:10.1155/2019/1523906 2019

  3. [3]

    Interference in Wireless Multi-Hop Ad-Hoc Networks and Its Effect on Network Capacity

    Hekmat, R., Van Mieghem, P. “Interference in Wireless Multi-Hop Ad-Hoc Networks and Its Effect on Network Capacity”. Wireless Networks vol 10, 389–399 (2004). https://doi.org/10.1023/B:WINE.0000028543.41559.ed

    work page arXiv 2004

  4. [4]

    Modelling and Simulation of Computer Networks and Systems: Methodologies and Applications

    M. S. Obaidat, F Zarai, P Nicopolitidis, “Modelling and Simulation of Computer Networks and Systems: Methodologies and Applications” Morgan Kaufman, Chapter 23, p.654 (2015)

    work page 2015

  5. [5]

    A survey on routing algorithms for wireless Ad-Hoc and mesh networks

    Eiman Alotaibi, Biswanath Mukherjee, “A survey on routing algorithms for wireless Ad-Hoc and mesh networks” Computer Networks, Volume 56, Issue 2, 2012, pp 940-965, https://doi.org/10.1016/j.comnet.2011.10.011

    work page doi:10.1016/j.comnet.2011.10.011 2012

  6. [6]

    A location aware transmission metric (ALARM) for multi-hop, multi-channel wireless mesh networks

    E. Al Otaibi, S. Roy, “A location aware transmission metric (ALARM) for multi-hop, multi-channel wireless mesh networks”, in Proceedings of the Wireless Communication and Networking Conference (WCNC’08), March 2008, pp. 2081–86

    work page 2008

  7. [7]

    Ad hoc routing protocols with multipoint relaying

    Géraud Allard, Philippe Jacquet, Laurent Viennot. Ad hoc routing protocols with multipoint relaying. 5 es rencontres francophones sur les Aspects Algorithmiques des Télécommunications (ALGOTEL), May 2003, Banyuls-sur- mer, France. inria-00471709

    work page 2003

  8. [8]

    A witness-aided transmission protocol for mobile ad hoc networks with unidirectional links

    I. Aron, S. Gupta, “A witness-aided transmission protocol for mobile ad hoc networks with unidirectional links”, in Proceedings of the First International Conference on Mobile Data Access, 1999, pp. 24–33

    work page 1999

Show all 21 references
  1. [9]

    SrcRR: A high throughput routing protocol for 802.11 mesh networks (DRAFT)

    Aguayo, Daniel, John Bicket, and Robert Morris. "SrcRR: A high throughput routing protocol for 802.11 mesh networks (DRAFT)." MIT, Tech. Rep (2005). Online: https://www.gta.ufrj.br/ensino/CPE825/2006/srcrr-draft.pdf

    2005

  2. [10]

    C Wahi1, S K Sonbhadra , “Mobile Ad Hoc Network Transmission Protocols: A Comparative Study’’ , International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.2, April 2012

    2012

  3. [11]

    Physical-layer Network Coding: Tutorial, survey, and beyond

    S Liew, S Zhang, "Physical-layer Network Coding: Tutorial, survey, and beyond." Physical Communication, Volume 6, pp 4-42 (2013)

    2013

  4. [12]

    Wireless Communications

    A. Goldsmith, "Wireless Communications” Cambridge University Press. p. 40-42 (2005)

    2005

  5. [13]

    T Riihonen, S Werner, and R Wichman ‘’ Comparison of Full-Duplex and Half-Duplex Modes with a Fixed Amplify- and-Forward Relay’’ in Proceedings of the Wireless Communication and Networking Conference (WCNC’0 9), March 2009

    2009

  6. [14]

    Antenna Engineering Handbook (2nd ed.)

    R. Johnson, "Antenna Engineering Handbook (2nd ed.)” (1984). New York, NY: McGraw-Hill, Inc. p. 1-12

    1984

  7. [15]

    The capacity of wireless networks,

    P. Gupta and P. R. Kumar, "The capacity of wireless networks," in IEEE Transactions on Information Theory, vol. 46, no. 2, pp. 388-404, March 2000, doi: 10.1109/18.825799

    doi

  8. [16]

    Universal Wireless Personal Communications

    R. Prasad, “Universal Wireless Personal Communications”, Artech House Publishers, 1998

    1998

  9. [17]

    The Alohanet - surfing for wireless data [History of Communications]

    M. Schwartz; N. Abramson ‘ “The Alohanet - surfing for wireless data [History of Communications]” in IEEE Communications Magazine, Volume: 47, Issue 12, Dec 2009

    2009

  10. [18]

    TDMA-based scheduling for multi-hop wireless sensor networks with 3-egress gateway linear topology

    L V Nguyen and N V Ha, “TDMA-based scheduling for multi-hop wireless sensor networks with 3-egress gateway linear topology” in Elsevier Internet of Things Journal, vol 14 (2021) https://doi.org/10.1016/j.iot.2021.100398

    2021

  11. [21]

    Comparison Studies for Different Shortest path Algorithms

    A. Tamimi, H. Abu-Ryash, “Comparison Studies for Different Shortest path Algorithms” International Journal of Computers and Applications, vol. 14, no. 8 (2015) DOI: 10.24297/ijct.v14i8.1857

    2015 doi

  12. [22]

    Wireless link scheduling under physical interference model,

    P. -J. Wan, O. Frieder, X. Jia, F. Yao, X. Xu and S. Tang, "Wireless link scheduling under physical interference model," 2011 Proceedings IEEE INFOCOM, Shanghai, China, 2011, pp. 838-845

    2011

  13. [23]

    Network Coding for Wireless Applications: A review

    Elom Worlanyo, “Network Coding for Wireless Applications: A review” (2016). Online: https://www.cse.wustl.edu/~jain/cse574-16/ftp/netcode.pdf

    2016

Pith tools

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