{"id":"9b6dba89-c8a5-42a8-b248-d45a0f8bb970","arxiv_id":"2411.14000","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"An integrated SUMO-BlockSim simulator shows that a retransmission-aware multi-connectivity scheme improves blockchain transaction throughput in V2X networks.","lead":"This paper links two simulators, SUMO for vehicle traffic and BlockSim for blockchain, to test how a new channel-switching rule called Enhanced MAX-SINR affects blockchain performance in satellite-assisted vehicle networks. It reports higher message and transaction throughput than a reinforcement learning baseline, especially when retransmissions succeed more often.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Single 50-second run per setting, with no seeds or confidence intervals in Section V, leaves the reported 18.71% throughput advantage over MARL statistically unsupported.","rationale":"Agree with the reader's weakest assumption: the handling of stochastic variability is the most load-bearing issue. Even a perfectly designed simulator yields a random variable as output; without repeated seeds the point estimate is just one draw. This is especially true here because Algorithm 1 uses random strategy selection and the channel model includes fading, and because transaction throughput depends on random block times. The reader's conditional verdict is appropriate: the paper's contributions — the integrated SUMO/BlockSim/V2X/satellite architecture, the released code, and the reference method — are real and worth publishing after the statistical gap is addressed. I therefore keep the verdict unchanged (still CONDITIONAL) rather than moving to ACCEPT or REJECT, because the concern is about missing evidence, not a demonstrated error. I also note the abstract's claim of evaluation in urban, suburban, and rural areas is not separately reported in the results, but this is secondary to the headline number.","tokens_in":7562,"tokens_out":5861,"duration_ms":58763,"concrete_test":"Using the released V2XBlockchain repository, run the N=300-vehicle scenario 10 times for both Enhanced MAX-SINR and MARL, with all Table I parameters fixed and only random seeds varied (SUMO route seed, PHY fading seed, and the Algorithm 1 strategy-selection seed). Report per-run transaction throughput, then the mean difference with a 95% confidence interval and a paired test. If the CI includes 0 or the 18.71% gap is not significant, the headline claim is not established; if the CI excludes 0, the single-run objection is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that Enhanced MAX-SINR raises blockchain transaction throughput by 18.71% over MARL (264.286 vs 222.62 TX/s at 300 vehicles). For this to be credible, the difference must reflect the method and not a single stochastic realization. Section V-A fixes simTime=50 s and Section V-B reports only point values; no random seeds, repeated runs, or confidence intervals are given. This is not a formality: Algorithm 1 intentionally uses 'strategy = random integer in {0,1,2,3}' for each failed retransmission, and the physical layer includes Rayleigh fading and log-normal shadowing, so the simulation is stochastic. One 50-second realization can easily differ by more than 41.7 TX/s if per-run variance is moderate, especially because the blockchain timeline itself has random block generation (Eq. (2)) and random block reception (Eq. (3)). The causal mechanism claimed in the abstract — that raising retransmission success rates boosts throughput — is also asserted rather than directly measured; no retransmission-success metric is reported. The released code is a real asset, but the headline number currently lacks the statistical support needed to distinguish signal from noise.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents a multi-layer simulation architecture that couples SUMO vehicular traffic with the BlockSim blockchain simulator and a satellite-assisted V2X connectivity layer. The authors implement and release an open-source V2X-blockchain simulator, address the synchronization of the two independent timelines (message delivery and block generation), and propose a reference multi-connectivity management method, Enhanced MAX-SINR, which adds four retransmission strategies to the conventional MAX-SINR selection. Evaluations in mixed urban/suburban/rural scenarios compare Enhanced MAX-SINR against MARL [9], MAX-SINR, and Random selection, reporting message throughput and transaction throughput for vehicle counts from 100 to 500. The headline result is 264.286 TX/s at 300 vehicles for Enhanced MAX-SINR, an 18.71% improvement over the MARL baseline's 222.62 TX/s.","tokens_in":7799,"tokens_out":5348,"duration_ms":48499,"significance":"The simulator integration and public code release are the strongest assets: combining SUMO and BlockSim with satellite-assisted V2X connectivity, and explicitly handling the two independent timelines, is useful and reusable for the community. The proposed Enhanced MAX-SINR is simple enough to serve as a benchmark and follows a plausible congestion-avoidance intuition. However, the quantitative contribution is currently supported only by single point estimates from a short stochastic simulation, and the causal mechanism asserted in the abstract is not directly measured. The contribution is therefore promising but not yet substantiated to the standard required for a headline performance claim.","major_comments":[{"comment":"The headline claim—an 18.71% transaction-throughput gain of Enhanced MAX-SINR over MARL at 300 vehicles (264.286 vs. 222.62 TX/s)—is based on a single 50-second simulation run per configuration, with no seeds, repetitions, confidence intervals, or variance estimates reported. This is a serious concern because the simulator is stochastic at several levels: Algorithm 1 (line 6) chooses retransmission strategies uniformly at random, the physical layer includes Rayleigh fading and log-normal shadowing, and block generation/reception times are exponential (Eqs. (2)–(3)). With simTime=50 s and TG=2.7 s, only roughly 18 block slots are generated per run, so the 41.7 TX/s difference could easily be within run-to-run variability. Please repeat each configuration over multiple independent seeds and report means with confidence intervals (or all per-run values), preferably with paired comparisons across methods.","section":"§V-B, Figs. 6–7; Table I"},{"comment":"The MARL baseline is taken from reference [9], which the paper describes as optimizing a non-blockchain utility function (throughput, delay, and power constraints), and no implementation details of the MARL policy used in this simulation are given (training steps, reward weights, architecture, seeds). The comparison therefore mixes two differences: the multi-connectivity management algorithm and the optimization objective. Either adapt the MARL reward to a blockchain-oriented metric, report its training setup and variance under the same conditions, or explicitly limit the claim to 'outperforms the published MARL policy of [9]' rather than 'outperforms MARL-based approaches' generally.","section":"§V-A; §V-B"},{"comment":"The paper's stated mechanism is that Enhanced MAX-SINR 'enhances the success rate of retransmitted messages' and that this raises blockchain transaction throughput, but no retransmission-success metric is reported anywhere in Section V. Without a measurement of retransmission success probability (or failure rate) per method, the data do not establish the causal link: the gain may come from better first-attempt selection, shorter delay, or fewer retransmissions overall. Please report retransmission-level statistics and, if possible, analyze throughput conditioned on retransmission success.","section":"Abstract; §IV; §V-B"},{"comment":"The 50-second simulation horizon and 2.7-second average block interval yield only about 18 block slots per run; with exponential block generation and block broadcast delays, a single run may contain few finalized blocks, making transaction throughput in Eq. (6) sensitive to the exact realization of forks and synchronization. The paper neither discusses burn-in/steady-state behavior nor reports the number of blocks actually generated per scenario. Please add this information and, ideally, longer-horizon results to show that the ordering among methods is stable.","section":"§V-A, Table I; Eq. (6)"}],"minor_comments":[{"comment":"The abstract contains the typo 'retransmission successfull rates', and the Fig. 1 caption reads 'The the multi-layer model'; both should be corrected.","section":"Abstract; Fig. 1"},{"comment":"Equation (5) is typeset as 'M EP LM' and does not convey the intended division; it should be written as a fraction, e.g., M_E / sum(L_M), with the symbols clearly defined.","section":"Eq. (5)"},{"comment":"In Table I, the unit for 'GLB: Block Gas Limit' is given as '(s)', which is not a gas unit; and Section V-A states vehicle densities of 'over 2000, 1000, and 400 vehicles per km respectively', where the unit should be clarified (per km² or per road km).","section":"Table I; §V-A"},{"comment":"Reference [8] lacks volume and page information, and reference [11] lacks a volume number; please complete the bibliographic details.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and the code release is a genuine asset. My main reservation is empirical: the paper's central performance claim needs proper statistical reporting before acceptance. The issues are fixable within revision—adding multiple seeds, confidence intervals, retransmission-success metrics, and a fairer MARL baseline—so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague, here's my read. The paper is an engineering integration effort: SUMO traffic, BlockSim blockchain, satellite-assisted multi-connectivity, and a small heuristic (Enhanced MAX-SINR) that adds four retransmission strategies on top of MAX-SINR. That combination is genuinely new to me. The code is on GitHub, which is a real asset; Table I lists most simulation parameters, so the environment is largely reproducible. Credit also goes to the authors for stating openly that the MARL baseline was optimized for non-blockchain objectives rather than overclaiming a fair comparison.\n\nThe soft spot is the one the stress-test flagged: the 18.71% throughput advantage rests on single runs. Section V-A sets simTime = 50 s, and Section V-B reports point values only. No seeds, no repeated runs, no confidence intervals. Algorithm 1 has explicit randomness (strategy = random integer in {0,1,2,3}), and the physical layer includes Rayleigh fading, log-normal shadowing, and exponential block generation/reception times. A 50-second run is short enough that stochastic variation could easily produce differences of the size reported. Until they add variance reporting, the claimed gain over MARL is plausible but unproven.\n\nA related gap: the paper states the causal mechanism—better retransmission success raises blockchain throughput—but never reports retransmission success rates. That metric is cheap to add and would directly test the hypothesis. Also, the MARL baseline from [9] is a self-citation; using one's own baseline is fine, but the exact MARL configuration and hyperparameters should be specified so the comparison is checkable. Minor point: the block-time experiment in Fig. 8 suffers the same single-run issue.\n\nOverall, this is a competent simulation study and a reasonable reference heuristic for blockchain-in-V2X testbeds. It will be most useful to researchers building on the released simulator. I would send it to peer review rather than desk-reject, because the integration and code are worth referee time, but I would expect heavy revision: multi-seed runs, confidence intervals, the retransmission-success metric, and clearer MARL details. If those are added, the paper becomes a solid contribution.\n\nRecommendation: conditional accept after the authors supply the missing statistical support and the directly measured mechanism.","headline":"A useful open-source V2X-blockchain simulator with a plausible retransmission heuristic; the headline throughput gain is statistically unsupported by single-run results.","tokens_in":8304,"tokens_out":2035,"would_cite":true,"duration_ms":20659,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that a retransmission-aware multi-connectivity rule raises blockchain transaction throughput in V2X networks by 18.71%.","keywords":["blockchain","V2X","Social Internet of Vehicles","multi-connectivity management","retransmission","satellite communication","performance evaluation","multi-agent reinforcement learning"],"falsifier":"Re-run the reported 300-vehicle urban scenario with 10 or more random seeds, varying traffic starting positions, channel fading realizations, and hash-power sampling, and compute confidence intervals for transaction throughput. If the intervals for Enhanced MAX-SINR and MARL overlap, the headline improvement is not established; if the ordering reverses, the central claim fails.","tokens_in":7362,"feed_emoji":"⛓️","tokens_out":8408,"duration_ms":67286,"temperature":0.7,"pith_summary":"This paper argues that in satellite-assisted vehicle-to-everything (V2X) networks where roadside units run a blockchain, the way vehicles select their radio connections determines how many vehicle messages become confirmed transactions. The authors propose a reference scheme, Enhanced MAX-SINR, that adds retransmission-aware choices to the standard maximum-SINR selection rule, and test it in a multi-layer simulator that couples SUMO traffic mobility, radio propagation including low-Earth-orbit satellite links, and the BlockSim blockchain simulator. At a vehicle count of 300, the scheme reaches 264.286 transactions per second, an 18.71% increase over a multi-agent reinforcement learning baseline that manages spectrum for non-blockchain goals. The paper's advertised takeaway is that boosting the success rate of retransmitted blockchain-related messages is the lever that improves blockchain performance, and that the released open-source simulator allows future intelligent management schemes to be developed and tested.","feed_headline":"Retransmission-aware channel selection lifts V2X blockchain throughput 18.71%","feed_subtitle":"Simulation with traffic, satellite links, and blockchain shows Enhanced MAX-SINR beats MARL at 264.286 TX/s.","key_machinery":"The load-bearing mechanism is Algorithm 1, the Enhanced MAX-SINR selection method. For each transmission attempt, a vehicle chooses a connection mode, transmit power, and sub-channel; on a failed attempt it randomly selects one of four recovery strategies, which lets it escape congested channels that plain MAX-SINR would re-enter because it only considers interference from the previous transmission. Around this, the three-layer simulator, combining SUMO for vehicle mobility, a radio layer with sub-channels and satellite links, and BlockSim for the RSU miner blockchain, runs on a single global clock. Messages arrive at RSUs with latencies set by the chosen connection mode; when the global clock matches the arrival time, the message is verified and added to the transaction pool, and when it matches exponentially distributed block generation and reception times, blocks are minted and broadcast. That arrangement is what lets the authors tie retransmission success to the number of confirmed transactions per second.","core_discovery":"The central discovery is that a multi-connectivity management rule explicitly tuned for retransmission success substantially improves blockchain throughput in an integrated V2X-blockchain simulation. When a vehicle's previous transmission fails, Enhanced MAX-SINR randomly picks among four strategies: switching to the best-SINR connection in a different connection mode, switching to a different sub-channel in the same mode, keeping the previous transmission selection, or reapplying plain MAX-SINR. Across urban, suburban, and rural traffic densities, the rule outperforms MARL, MAX-SINR, and random selection, and the paper reports the highest transaction throughput at 300 vehicles, 264.286 TX/s versus 222.62 TX/s for MARL. The paper also identifies system-level implementation challenges, such as two simulation timelines across the transportation, connection, and consensus layers, and resolves them with a global clock that dictates when messages enter the transaction pool and when blocks are generated and synchronized.","pith_inferences":["The 18.71% gain is against a MARL agent that optimizes non-blockchain throughput and delay; a stricter comparison would pit Enhanced MAX-SINR against a method also optimized for retransmission success, to separate the benefit of retransmission awareness from the effect of simply having more choices after a failure.","Because the evaluation uses one run per setting, the paper's strongest product is the simulator itself rather than the specific throughput ordering; multiple seeds would turn the reported numbers into a statistically grounded comparison.","The paper leaves implicit that the same global-clock architecture could be used to localize the dominant bottleneck, whether retransmission failures or consensus parameters, by varying one layer while holding the others fixed.","A testable extension the authors do not run is to feed retransmission-failure statistics into the MARL reward function so agents learn to minimize the blockchain-level cost, which could combine MARL's adaptivity with the reference scheme's retransmission awareness."],"forward_implications":["Transaction throughput saturates around 300 vehicles for every tested method, suggesting that in dense traffic the blockchain consensus layer, not radio access, becomes the binding constraint on confirmed transactions.","Multi-connectivity management for blockchain-based V2X should be designed around retransmission success, because a failed retransmission removes a message from the pool that would have become a transaction.","Adjusting the block generation time trades throughput against fork risk: shorter block intervals can raise throughput but also increase forks and unconfirmed transactions, and Enhanced MAX-SINR stays highest across all tested block intervals.","The open-source multi-layer simulator provides a testbed on which future intelligent management methods can be evaluated in an environment that includes traffic, radio, and consensus together."],"supporting_citations":[{"why":"Supplies the MARL baseline that the proposed method is compared against; without it, the reported 18.71% improvement over a non-blockchain-optimized agent would not be defined.","marker":"[9]"},{"why":"Establishes that message retransmission is a key obstacle to blockchain transaction throughput, the motivation for the retransmission-aware design.","marker":"[16]"},{"why":"Provides BlockSim, the blockchain simulator used to model RSU miners, transaction pools, block generation, and synchronization.","marker":"[17]"},{"why":"Supplies real Starlink satellite orbital data used in the satellite-assisted connectivity environment.","marker":"[19]"},{"why":"Defines the TR 38.886 pathloss model used for vehicle-to-infrastructure and vehicle-to-vehicle radio links.","marker":"[20]"},{"why":"Defines the TR 38.821 pathloss model used for vehicle-to-satellite links.","marker":"[21]"},{"why":"Provides the values for average block generation time and block receive time used as blockchain simulation parameters.","marker":"[22]"},{"why":"Supplies the definition of transaction throughput as the metric for blockchain performance evaluation.","marker":"[23]"}],"fun_headline_variants":["Retransmission-aware channel selection lifts V2X blockchain throughput","Enhanced MAX-SINR rule boosts V2X blockchain to 264.286 TX/s","V2X blockchain sim: retransmission success rate drives throughput","Multi-connectivity rule outperforms MARL in V2X blockchain","Retransmission-focused management lifts blockchain TX rate in V2X"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The evaluation rests on treating one 50-second simulation run per setting as representative: the paper reports point values without seeds, repeated runs, or confidence intervals, so the claimed 18.71% gain could be within run-to-run variability.","fun_headline_variants_meta":{"raw":{"variants":["Retransmission-aware channel selection lifts V2X blockchain throughput","Enhanced MAX-SINR rule boosts V2X blockchain to 264.286 TX/s","V2X blockchain sim: retransmission success rate drives throughput","Multi-connectivity rule outperforms MARL in V2X blockchain","Retransmission-focused management lifts blockchain TX rate in V2X"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000364,"raw_usage":{"total_tokens":1939,"prompt_tokens":903,"completion_tokens":1036,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":519,"completion_tokens_details":{"reasoning_tokens":940}},"tokens_in":519,"tokens_out":1036,"duration_ms":8327,"temperature":1.0,"reasoning_tokens":940,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:38:15.218350+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the reported 300-vehicle urban scenario with 10 or more random seeds, varying traffic starting positions, channel fading realizations, and hash-power sampling, and compute confidence intervals for transaction throughput. If the intervals for Enhanced MAX-SINR and MARL overlap, the headline improvement is not established; if the ordering reverses, the central claim fails.","supporting_citations":[{"cited_title":"Multi-Agent Deep Reinforcement Learning for Spectrum Management in V2X with Social Roles,","cited_arxiv_id":null,"evidence_quote":"Supplies the MARL baseline that the proposed method is compared against; without it, the reported 18.71% improvement over a non-blockchain-optimized agent would not be defined."},{"cited_title":"A Blockchain-Enabled Vehicular Edge Computing Framework for Secure Performance-oriented V2X Service Delivery,","cited_arxiv_id":null,"evidence_quote":"Establishes that message retransmission is a key obstacle to blockchain transaction throughput, the motivation for the retransmission-aware design."},{"cited_title":"Blocksim: An extensible simulation tool for blockchain systems,","cited_arxiv_id":null,"evidence_quote":"Provides BlockSim, the blockchain simulator used to model RSU miners, transaction pools, block generation, and synchronization."},{"cited_title":"Starlink Satellite Elements,","cited_arxiv_id":null,"evidence_quote":"Supplies real Starlink satellite orbital data used in the satellite-assisted connectivity environment."},{"cited_title":"V2X Services based on NR; User Equipment (UE) radio transmission and reception; (Release 16),","cited_arxiv_id":null,"evidence_quote":"Defines the TR 38.886 pathloss model used for vehicle-to-infrastructure and vehicle-to-vehicle radio links."},{"cited_title":"3GPP TR 38.821: Solutions for NR to support non-terrestrial networks (NTN),","cited_arxiv_id":null,"evidence_quote":"Defines the TR 38.821 pathloss model used for vehicle-to-satellite links."},{"cited_title":"An Effec- tive Framework of Private Ethereum Blockchain Networks for Smart Grid,","cited_arxiv_id":null,"evidence_quote":"Provides the values for average block generation time and block receive time used as blockchain simulation parameters."},{"cited_title":"Performance evaluation of blockchain systems: A systematic survey,","cited_arxiv_id":null,"evidence_quote":"Supplies the definition of transaction throughput as the metric for blockchain performance evaluation."}],"review_version":1}