Predictive-CSM: Lightweight Fragment Security for 6LoWPAN IoT Networks

Somayeh Sobati-M

REVIEW 4 major objections 6 minor 40 references

Predictive-CSM: Lightweight Fragment Security for 6LoWPAN IoT Networks

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A two-layer filter of trust scoring and chained hashes can stop 6LoWPAN fragmentation attacks while keeping delivery above 97%.

desk verdict The trust-plus-hash combination is sensible, but the published protocol never validates FRAG1 before buffer admission, so the central defense claim doesn't follow from the paper as written. read the letter →

arxiv 2506.01767 v1 pith:P3RC4RZC submitted 2025-06-02 cs.CR

Somayeh Sobati-M This is my paper

classification cs.CR

keywords 6LoWPANfragmentationattacksIoTsecuritytrustmanagementhashchainingHMACdenialofserviceContiki-NG

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

Predictive-CSM is a security add-on for 6LoWPAN networks that aims to stop fragmentation-based denial-of-service attacks before they exhaust the tiny reassembly buffers of constrained devices. The paper argues that a lightweight behavioral trust score, updated by an exponential moving average, plus a chained HMAC signature on each fragment, can filter malicious fragments at the adaptation layer without the energy cost of per-fragment MAC schemes like SecuPAN. In Cooja simulations with a star network of Tmote Sky-class motes, the scheme reports 97.4–99.2% packet delivery under five attack types, attack detection within 4.4–7.0 seconds, and average power of 0.33–0.35 mW versus 0.41–0.61 mW for SecuPAN. If correct, this would give resource-constrained IoT devices a deployable defense against buffer-reservation, flooding, replay, burst, and late-phase fragment attacks at near-baseline energy cost.

What carries the argument

The load-bearing mechanism is the pairing of a predictive trust score with an incremental hash chain, evaluated sequentially before a fragment is admitted to the reassembly buffer. The trust score is an exponential moving average over observed fragment outcomes, so a sender's reputation decays quickly on invalid fragments; the hash chain is a one-way link from seed $H_0 = \mathrm{HMAC}(K, d_0 \| \mathrm{nonce})$ through $H_i = \mathrm{HMAC}(K, H_{i-1} \| d_i)$, which lets the receiver verify each fragment's integrity and position in the sequence without holding the whole packet. Together the two checks give early rejection: the trust engine filters behaviorally suspicious sources, and the hash validator catches forged or misordered content even from a source that still has a high trust score.

What would settle it

Take a fresh Predictive-CSM receiver and replay a previously captured, cryptographically valid fragment chain from a legitimate sender without any other misbehavior. If the receiver admits the replayed fragments into reassembly and a genuine packet is lost before the sender's trust score drops below $\theta$, the claimed replay and buffer-reservation defense fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that fragment-level attacks on 6LoWPAN can be countered by a two-layer defense that lives in the adaptation layer rather than the routing layer or a heavyweight crypto layer. The first layer, the Predictive Trust Engine, scores each neighbor with an exponential moving average $T_i(t) = \lambda T_i(t-1) + (1-\lambda) O_i(t)$ and drops fragments from nodes whose score falls below a threshold $\theta$. The second layer, the Fragment Signature Validator, requires every fragment to carry a chained hash $H_i = \mathrm{HMAC}(K, H_{i-1} \| d_i)$, with a seed $H_0$ that includes a nonce, so tampered or reordered fragments fail before entering reassembly. The paper reports that this combination preserves packet delivery under early FRAG1 injection, full-fragment flooding, header-replay cloning, burst injection, and late-phase injection, while using less power than SecuPAN in every tested scenario.

Load-bearing premise

The argument assumes every sender and receiver already share a secret key $K$ for the HMAC chain and that the nonce in the seed hash stays fresh, but the paper does not specify how keys are established, how nonces are synchronized or checked, or how a replayed but cryptographically valid fragment chain is rejected before the trust score falls.

Editorial extensions

If this is right

If the reported results hold, fragment-level defenses no longer require per-fragment MACs with synchronized nonce state; the chained hash gives incremental integrity at lower energy cost.
An adaptive trust threshold means the same node can tolerate transient channel errors from historically reliable senders while still blacklisting persistent misbehavior, reducing false drops compared with strict cryptographic validation.
The scheme claims detection latencies of 4.4–7.0 seconds across all five attack types, which is fast enough to prevent reassembly-buffer exhaustion from evolving into full denial of service on Class 1 devices.
Because the extended headers are backward-compatible, Predictive-CSM could be deployed incrementally on existing 6LoWPAN stacks without changing the core standard.

Reading between the lines

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

The scheme's replay resistance rests on a nonce in the seed hash, but no nonce-freshness check is specified at the receiver; a captured valid fragment chain could plausibly be replayed until behavioral scoring lowers the sender's trust, so I would read the header-replay results as coverage against unauthenticated forgery rather than replay of legitimate captures.
The evaluation uses a 10-node star topology with a single attacker; the trust model and buffer dynamics would likely behave differently in a multi-hop mesh with Route-Over reassembly at every hop, where intermediate nodes could amplify or obscure trust signals.
A natural testable extension is to feed the trust engine with physical-layer features, such as received signal strength or arrival-time jitter, which the paper lists as future work; that could shorten the 4.4–7.0 second detection window for timing-offset attacks.
The buffer-availability model suggests that the scheme's main benefit under flooding is not cryptographic strength but early admission control; the same trust engine could be paired with any cheap integrity check and still preserve delivery, so the trust engine is likely the dominant contributor to the PDR gains.

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.

Desk Editor's Note

read the letter

Short version: this paper pairs an EWMA trust score with a chained-HMAC fragment check at the 6LoWPAN adaptation layer, and that combination is reasonable. But the central claim—that malicious fragments are caught before they occupy memory—is not supported by the protocol as written. The stress-test note holds up on reading. Algorithm 3 handles FRAG1 separately, checks only the trust score, then initializes the reassembly buffer and proceeds to signature validation. Since Algorithm 1 initializes unknown nodes at Tn = 0.5, above the chosen threshold θ = 0.3, a fresh attacker's first FRAG1 is admitted without any cryptographic check. On top of that, the seed hash H0 and nonce from Eq. 4 never appear in the sender/receiver algorithms, so the receiver cannot validate the first fragment even if it wanted to. The replay and buffer-reservation defenses therefore do not follow from the published protocol.

Credit where it's due: the paper clearly identifies a real gap—routing-layer CSM does not touch fragment-level attacks, and SecuPAN's per-fragment MAC is heavy for Class 1 devices. The five-scenario evaluation is a reasonable attempt, and the power numbers (0.33–0.35 mW vs 0.41–0.61 mW for SecuPAN) are plausible for a hash-plus-trust scheme. The analytical trust model is standard but correctly derived.

The soft spots beyond the FRAG1 hole: Complete Flooding is missing from the PDR and drop tables (Tables 3 and 4), despite being listed as a scenario. The default λ and θ are chosen from the same attack simulations rather than a held-out set, so the headline numbers are tuned, not predicted. There are no error bars on the 15-run averages, and no code or data release, so reproducibility is weak.

None of this kills the underlying idea. Reordering validation before buffer admission and transmitting the seed hash would fix the main protocol gap. But as written, the paper overclaims what it shows. I would send it to review—the subfield needs a lightweight alternative to SecuPAN, and the flaws are fixable—but I would expect major revision. For my own work, I would not cite it as evidence until the protocol and data are straightened out.

Referee Report

4 major / 6 minor

Summary. The paper proposes Predictive-CSM, a lightweight security framework for 6LoWPAN fragmentation attacks that combines a Predictive Trust Engine (PTE), which maintains an exponentially weighted moving-average trust score per neighbor, with a Fragment Signature Validator (FSV), which checks chained HMAC tags on fragments. The system is evaluated in Contiki-NG/Cooja simulations against five attack scenarios (early FRAG1 injection, complete fragment flooding, header-replay cloning, burst injection, and late-phase injection), reporting near-baseline power consumption (0.33–0.35 mW), high packet delivery ratios (97.4–99.2%), low fragment drop rates, and detection latencies of 4.4–7.0 seconds compared with vanilla 6LoWPAN, CSM, and SecuPAN. The paper also presents a sensitivity analysis of the trust parameters λ and θ, an analytical model of trust dynamics, and a buffer-availability model.

Significance. If the results are correct, Predictive-CSM would be a meaningful step toward practical fragment-level security for constrained IoT devices, offering a lower-cost alternative to per-fragment MAC schemes like SecuPAN while addressing buffer-reservation and replay-style attacks at the 6LoWPAN adaptation layer. The combination of behavioral trust and hash chaining is a sensible design, and the use of Contiki-NG/Cooja with 15-run averages is a reasonable evaluation approach for this class of protocols. However, the paper's central claims are currently undermined by missing results for one of the five stated attacks, an implementation-level gap in the FRAG1 admission path, an undefined seed-hash mechanism, and an in-sample parameter-selection procedure. These issues need to be resolved before the performance numbers can be taken as reliable evidence.

major comments (4)

[§5.3, Table 3, Table 4, Table 5] Complete Fragment Flooding is defined as the second attack scenario in Section 5.3, but it is absent from the PDR table (Table 3), the fragment drop table (Table 4), and the detection-latency table (Table 5); only power consumption for this scenario appears in Table 2. Consequently, the abstract's claim that Predictive-CSM preserved delivery 'across all scenarios' and the summary statistics '97.4–99.2% PDR' and '4.4–7.0 s' do not cover a stated attack. Please report PDR, drop rate, and detection latency for the Complete Flooding scenario, or explicitly justify its omission.
[Algorithm 3, Algorithm 1, §3.4] The protocol as written admits FRAG1 into the reassembly buffer before any FSV check. Algorithm 3 (lines 1–6) handles FRAG1 by checking only the trust threshold, then executes 'Store trust score for session and initialize reassembly' and 'return Proceed to signature validation'; no HMAC computation is performed for FRAG1. Since Algorithm 1 initializes an unknown node's trust to Tn = 0.5, which is above the configured threshold θ = 0.3, a fresh attacker's first FRAG1 passes the PTE and occupies a buffer slot. Section 7 states that blacklisting requires 3–4 invalid events, so under a burst attack (up to 6 FRAG1/s) several unauthenticated fragments will occupy memory before the trust score drops. This directly contradicts the abstract's claim that malicious fragments are caught 'before they can occupy memory'.
[Eq. (4), Algorithms 4–5, §4.1–4.2] The FSV cannot verify the first fragment because the seed hash and nonce are not transmitted. Equation (4) defines H0 = HMAC(K, d0 || nonce), but Algorithm 4 attaches only the trust score and the computed Hi to the header, with no explicit nonce or H0 field. On the receiver side, Algorithm 5 (and Algorithm 3 for non-FRAG1 fragments) computes Hexpected = HMAC(K, Hprev || payload), but Hprev is undefined for the first fragment. Unless the nonce and H0 are included in the extended header—which the pseudocode and message-format description do not state—the receiver cannot recompute the expected seed signature, so replay of captured FRAG1 headers cannot be countered by the hash chain. Please clarify the header format and specify how the receiver obtains the nonce and initial hash.
[§6.5, Table 6] The default configuration (λ = 0.9, θ = 0.3) is selected by comparing seven parameter combinations on the same attack simulations, and the paper then presents the corresponding PDR and detection latency as the system's result in Tables 3 and 5. There is no held-out validation set or cross-validation, so the headline performance figures are in-sample optima and the claim that these parameters are suitable for 'real-world IoT deployments' is not supported. Please evaluate the chosen configuration on an independent validation set or report the full performance distribution across parameter settings so that the selection bias can be assessed.

minor comments (6)

[§6.1–6.4] The paper reports averages over 15 independent runs without any error bars, standard deviations, or confidence intervals; for differences as small as 0.3% PDR between Predictive-CSM and SecuPAN in the no-attack case, this makes the comparison statistically unverifiable.
[§6.5, Table 6] The metric 'False Positives (%)' used in the sensitivity analysis is never defined; please clarify whether it refers to legitimate fragments dropped or legitimate packets lost, and how it was computed.
[Table 3] The column 'Gain vs CSM' lists percentage-point differences (e.g., 97.4 vs. 77.2 is labeled 20.2%) but the label '%' implies a relative improvement; please state which convention is used.
[Table 7, §7] The symbol λ is used both as the trust forgetting factor and, in the buffer model, as the arrival rate of valid fragments; the table distinguishes 'λ (buﬀer)' but the surrounding text does not, which is confusing.
[References] Several references are duplicated (e.g., [4] and [5]; [12] and [13]; [15] and [16]; [21] and [22]), and the bibliography should be deduplicated.
[General] There are many typographical and OCR-style artifacts (e.g., '6LoWPA N' in the title, 'a nd', 'ana lysis', 'eﬃciency') that should be corrected in a careful revision.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline performance is in-sample tuning: Section 6.5 selects the default (lambda, theta) on the same attack simulations whose PDR and latency are then reported as Predictive-CSM's result.

fitted input called prediction [Section 6.5, Table 6 (Parameter Sensitivity Analysis)]
"The chosen θ = 0.3 optimized both security and tolerance, with 98.9% PDR and 5.1-second median detection latency. ... Key Insight: As shown in Table 6, the default configuration (λ = 0.9, θ = 0.3) achieved optimal balance across all metrics, validating our design choices for real-world IoT deployments."

The default configuration is selected by comparing seven (lambda, theta) combinations on the same attack simulations that generate the paper's headline numbers. Table 6's Default row (98.9% PDR under attack, 5.1 s detection latency) is then presented in Tables 3 and 5 as Predictive-CSM's performance under the same attack families, for example 99.0% PDR for Early FRAG1, 98.9% for Header Replay, 97.4% for Burst Injection, and 5.1 s latency for Early FRAG1. The reported delivery and latency are therefore not independent predictions for a fixed configuration; they are the values corresponding to the configuration chosen to optimize those same metrics.

full rationale

Apart from the in-sample parameter selection, I do not find a strict circular derivation in this paper. No load-bearing claim rests on a self-citation: CSM and SecuPAN are prior external works, and the chained-HMAC construction is a standard one. The notable objections to the protocol—Algorithm 3 admits FRAG1 to reassembly state before signature validation, and Algorithms 4-5 never transmit H0 or the nonce to the receiver—are correctness gaps in the protocol as written, not cases where an output equals an input by definition. The one defensible circularity is the parameter-tuning pattern: Section 6.5 chooses lambda = 0.9 and theta = 0.3 using the same attack scenarios and metrics that are later reported as Predictive-CSM's headline PDR and detection latency, so the central quantitative claims are partly fitted rather than independently predicted. That supports a score of 6 rather than a higher score, because the simulations still measure a concrete protocol on specified scenarios; they are not a pure tautology.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim rests on two tuned parameters (lambda and theta), a shared-key assumption for HMAC, and a stated attacker model. The buffer availability formula is a rough analytical approximation, not a hardware measurement. No new physical entities are introduced.

free parameters (2)

lambda (forgetting factor) = 0.9
Chosen as the default after sensitivity analysis on the same attack simulations (Section 6.5.1, Table 6); not derived from first principles.
theta (trust threshold) = 0.3
Chosen as the default after sensitivity analysis on the same attack simulations (Section 6.5.2, Table 6); not independently justified.

assumptions (4)

domain assumption Each sender and receiver share a secret key K for HMAC computation, with no key-establishment or nonce-synchronization specified.
Algorithms 2, 4, and 5 invoke HMAC(K, ...) with a shared key; Section 4.1 uses a nonce in the seed hash but the receiver-side freshness check is not described.
domain assumption The attacker is an external node without access to valid cryptographic keys.
Section 5.3 adversarial model; this excludes insider attacks and limits the threat model.
domain assumption Malicious behavior produces observable deviations in fragment frequency, sequence order, timing, or payload plausibility that the PTE can score.
Section 3.2 Algorithm 1 lists these features; if an attacker mimics legitimate patterns, trust scores will not drop.
domain assumption The buffer occupancy model rho = min(1, (lambda + A)/(B * tau)) approximates reassembly buffer behavior.
Section 7.3 Eq. (6); a simplified M/D/1-like formula with no derivation, and parameters B and tau are not measured.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Predictive-CSM: Lightweight Fragment Security for 6LoWPAN IoT Networks." pith.science (2026). https://pith.science/paper/P3RC4RZC

@misc{pith2026250601767,
  author       = {Pith},
  title        = {Pith review of: Predictive-CSM: Lightweight Fragment Security for 6LoWPAN IoT Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P3RC4RZC}},
  note         = {Machine review of arXiv:2506.01767}
}

read the original abstract

Fragmentation is a routine part of communication in 6LoWPAN-based IoT networks, designed to accommodate small frame sizes on constrained wireless links. However, this process introduces a critical vulnerability fragments are typically stored and processed before their legitimacy is confirmed, allowing attackers to exploit this gap with minimal effort. In this work, we explore a defense strategy that takes a more adaptive, behavior-aware approach to this problem. Our system, called Predictive-CSM, introduces a combination of two lightweight mechanisms. The first tracks how each node behaves over time, rewarding consistent and successful interactions while quickly penalizing suspicious or failing patterns. The second checks the integrity of packet fragments using a chained hash, allowing incomplete or manipulated sequences to be caught early, before they can occupy memory or waste processing time. We put this system to the test using a set of targeted attack simulations, including early fragment injection, replayed headers, and flooding with fake data. Across all scenarios, Predictive CSM preserved network delivery and maintained energy efficiency, even under pressure. Rather than relying on heavyweight cryptography or rigid filters, this approach allows constrained de vices to adapt their defenses in real time based on what they observe, not just what they're told. In that way, it offers a step forward for securing fragmented communication in real world IoT systems

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

40 extracted references · 39 canonical work pages

[1]

Journal of Cloud Computing 12(1):42

Aaqib M, Ali A, Chen L, Nibouche O (2023) Iot trust and reputatio n: a survey and taxonomy. Journal of Cloud Computing 12(1):42

work page 2023
[2]

Ad Hoc Networks

Abbasi M, Al-Anbagi I (2019) Mitigation of fragmentation-base d dos attacks in 6lowpan networks using early drop mechanisms. Ad Hoc Networks

work page 2019
[3]

Applied Sciences 12(17):8605

Ahmad R, Wazirali R, Abu-Ain T, Almohamad TA (2022) Adaptive tru st-based framework for securing and reducing cost in low-cost 6lowpan wireless sensor n etworks. Applied Sciences 12(17):8605

work page 2022
[5]

Sen sors 22(24), URL https://www.mdpi.com/1424-8220/22/24/9825

Alyami S, Alharbi R, Azzedin F (2022) Fragmentation attacks an d countermeasures on 6lowpan internet of things networks: Survey and simulation. Sen sors 22(24), URL https://www.mdpi.com/1424-8220/22/24/9825

work page 2022
[6]

Bao F, Chen IR (2012) Dynamic trust management for internet o f things applications. In: Proceedings of the 2012 International Workshop on Self-Awa re Internet of Things, Association for Computing Machinery, New York, NY, USA, Self-IoT ’12, p 1–6, DOI 10.1145/2378023.2378025, URL https://doi.org/10.1145/2378023.2378025 22 Somayeh Sobati-Moghadam

work page arXiv 2012
[7]

Wireless Personal Communications 113:1–17

Elgendy H, Aly SM, Nabil A (2020) Trust-based model for secure routing in wireless sensor networks. Wireless Personal Communications 113:1–17

work page 2020
[8]

Future Internet 16(10), DOI 10.3390/ﬁ16100368, URL https://www.mdpi.com/1999-5903/16/10/368

Fatima M, Rehman O, Rahman IMH, Ajmal A, Park SJ (2024) Towar ds ensemble feature se- lection for lightweight intrusion detection in resource-constrained iot devices. Future Internet 16(10), DOI 10.3390/ﬁ16100368, URL https://www.mdpi.com/1999-5903/16/10/368

work page 2024
[9]

Journal of Systems Architecture

Firoozi F, et al (2021) A trust-based method for secure comm unication in internet of things using edge and fog computing. Journal of Systems Architecture

work page 2021

Show all 40 references

[10]

Ad Hoc Networks

Ghosh A, et al (2018) Trustlite: Lightweight trust managemen t scheme for resource- constrained iot devices. Ad Hoc Networks

2018
[11]

IEEE Internet of Things Journal 8(6):4022–4037

Ghubaish A, Al-Rubaye M, Tsourdos A (2021) A comprehensive s urvey of trust management in iot. IEEE Internet of Things Journal 8(6):4022–4037

2021
[12]

Computer Networks 149:113–133

Glissa G, Rachedi A (2019) Secure and eﬃcient data transmissio n for constrained iot devices: A survey. Computer Networks 149:113–133

2019
[13]

Computer Networks 149:113–133

Glissa G, Rachedi A, Meddeb A (2019) Secure and eﬃcient data t ransmission for constrained iot devices: A survey. Computer Networks 149:113–133

2019
[14]

Computers, Materials, Continua 83(3):5539–5559, URL http://www.techscience.com/cmc/v83n3/60981

Hongliang Tian ML (2025) A lightweight iot data security sharing sc heme based on attribute- based encryption and blockchain. Computers, Materials, Continua 83(3):5539–5559, URL http://www.techscience.com/cmc/v83n3/60981

2025
[15]

In: Proceedings of the Eighth A CM Conference on Data and Application Security and Privacy, Association for Computing Mac hinery, New York, NY, USA, CODASPY ’18, p 307–318

Hossain M, Karim Y, Hasan R (2018) Secupan: A security scheme to mitigate fragmentation- based network attacks in 6lowpan. In: Proceedings of the Eighth A CM Conference on Data and Application Security and Privacy, Association for Computing Mac hinery, New York, NY, USA, CODAS...

2018
[16]

Proc ACM CODASPY

Hossain M, et al (2018) Secupan: A security scheme to mitigate fragmentation-based network attacks in 6lowpan. Proc ACM CODASPY

2018
[17]

Association for Computing Machin- ery, New York, NY, USA, WiSec ’13, p 55–66, DOI 10.1145/2462096.2 462107, URL https://doi.org/10.1145/2462096.2462107

Hummen R, Hiller J, Wirtz H, Henze M, Shafagh H, Wehrle K (2013) 6 lowpan fragmentation attacks and mitigation mechanisms. Association for Computing Machin- ery, New York, NY, USA, WiSec ’13, p 55–66, DOI 10.1145/2462096.2 462107, URL https://doi.org/10.1145/2462096.2462107

2013
[18]

Journal of Systems Architecture

Jangra A, et al (2021) An adaptive trust-aware buﬀer manag ement strategy for resource- constrained iot devices. Journal of Systems Architecture

2021
[19]

International Journal of Information Security 22:245–257

Khan M, Hayat A (2023) Real-time anomaly detection in iot networ ks. International Journal of Information Security 22:245–257

2023
[20]

Comp uters, Materials and Continua 78(1):31–63

Kumar S, Kumar D, Dangi R, Choudhary G, Dragoni N, You I (20 24) A review of lightweight security and privacy for resource-constrained iot devices. Comp uters, Materials and Continua 78(1):31–63
[21]

Journal of Network and Computer Applications

Lakshmi V, et al (2021) A lightweight data authentication mode l for iot networks using incremental hmac veriﬁcation. Journal of Network and Computer Applications

2021
[22]

J Network and Computer Applications

Lakshmi V, et al (2021) Lightweight hmac veriﬁcation for 6lowp an fragments. J Network and Computer Applications

2021
[23]

IEEE Transactions on Information Forensics and Security 17:456–469

Li J, Wang L (2022) A lightweight cryptographic protocol for 6 lowpan. IEEE Transactions on Information Forensics and Security 17:456–469

2022
[24]

IEEE Internet of Things Journal 10(7):5898–5922, DOI 10.1109/JIOT.2023.3237893

Liu Y, Wang J, Yan Z, Wan Z, J¨ antti R (2023) A survey on blockc hain-based trust man- agement for internet of things. IEEE Internet of Things Journal 10(7):5898–5922, DOI 10.1109/JIOT.2023.3237893

2023
[25]

Journal of Systems Architecture 131:102590

Mansoor A, Ali Q (2023) Integrating blockchain for data integr ity in iot. Journal of Systems Architecture 131:102590

2023
[26]

Computer Networks Predictive-CSM: Lightweight Fragment Security for 6LoWPA N IoT Networks 23

Mendoza A, Jara AJ, Skarmeta AF (2015) Adaptive trust mana gement for 6lowpan routing. Computer Networks Predictive-CSM: Lightweight Fragment Security for 6LoWPA N IoT Networks 23

2015
[27]

URL https://arxiv.org/abs/2311.11705, 2311.11705

Okporokpo O, Olajide F, Ajienka N, Ma X (2023) Trust-based ap proaches towards enhanc- ing iot security: A systematic literature review. URL https://arxiv.org/abs/2311.11705, 2311.11705

2023 arXiv
[28]

In: Proceedings of 31st IEEE Conference on Lo cal Computer Networks, IEEE, pp 641–648

Osterlind F, Dunkels A, Eriksson J, Finne N, Voigt T (2006) Cross -level sensor network sim- ulation with cooja. In: Proceedings of 31st IEEE Conference on Lo cal Computer Networks, IEEE, pp 641–648

2006
[29]

IEEE communi- cations surveys , tutorials 15(3):1389–1406

Palattella MR, Accettura N, Vilajosana X, Watteyne T, Grieco LA , Boggia G, Dohler M (2013) Standardized protocol stack for the internet of (import ant) things. IEEE communi- cations surveys , tutorials 15(3):1389–1406

2013
[30]

URL https://arxiv.org/abs/2006.00310, 2006.00310

Raoof A, Lung CH, Matrawy A (2020) Introducing network cod ing to rpl: The chained secure mode (csm). URL https://arxiv.org/abs/2006.00310, 2006.00310

2020 arXiv
[31]

Journal of Info rmation and Intelligence 2(6):455–513

Sasi T, Lashkari AH, Lu R, Xiong P, Iqbal S (2024) A comprehe nsive survey on iot attacks: Taxonomy, detection mechanisms and challenges. Journal of Info rmation and Intelligence 2(6):455–513

2024
[32]

Journal of Network and Computer Applications 178:102918

Sharma A, Gupta R (2021) An adaptive trust management syst em for iot devices. Journal of Network and Computer Applications 178:102918

2021
[33]

IEEE Access 8:163077–163090, DOI 10.1109/ACCESS.2020.3020129

Su B, Du C, Huan J (2020) Trusted opportunistic routing based on node trust model. IEEE Access 8:163077–163090, DOI 10.1109/ACCESS.2020.3020129

2020
[34]

Sensors 22(6)

Sultana S, Khan A (2022) Trust-aware rpl for 6lowpan securit y. Sensors 22(6)

2022
[35]

Sensors 22(6):2295

Sultana S, Khan A, Aslam N (2022) An eﬃcient trust-aware rout ing protocol for securing 6lowpan in iot networks. Sensors 22(6):2295

2022
[36]

Team CN (2018) Contiki-ng: The os for next generation iot devic es URL https://contiki-ng.org

2018
[37]

https://rfc-editor.org/rfc/rfc8930.txt

Thubert P (2020) Rfc 8930 - on forwarding 6lowpan fragments over a multi-hop ipv6 network. https://rfc-editor.org/rfc/rfc8930.txt

2020
[38]

IEEE Internet of Things Journal

Wang J, Liu H, Zhang F (2020) A blockchain-based lightweight dat a authentication protocol for 6lowpan. IEEE Internet of Things Journal

2020
[39]

In: Proceedings of IEEE ICC

Xie L, Wang C, Zhou W (2020) A secure fragmentation mechanism for low-power networks. In: Proceedings of IEEE ICC

2020
[40]

Journal of Sensors

Yousuf A, Ismail ASKP (2017) Trust management in wireless sen sor networks: an overview. Journal of Sensors

2017
[41]

Future Generation Computer Systems 128:103–113

Zhang Y, Chen X (2022) A hybrid trust-based framework for io t security. Future Generation Computer Systems 128:103–113

2022

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Journal of Cloud Computing 12(1):42

Aaqib M, Ali A, Chen L, Nibouche O (2023) Iot trust and reputatio n: a survey and taxonomy. Journal of Cloud Computing 12(1):42

work page 2023

[2] [2]

Ad Hoc Networks

Abbasi M, Al-Anbagi I (2019) Mitigation of fragmentation-base d dos attacks in 6lowpan networks using early drop mechanisms. Ad Hoc Networks

work page 2019

[3] [3]

Applied Sciences 12(17):8605

Ahmad R, Wazirali R, Abu-Ain T, Almohamad TA (2022) Adaptive tru st-based framework for securing and reducing cost in low-cost 6lowpan wireless sensor n etworks. Applied Sciences 12(17):8605

work page 2022

[4] [5]

Sen sors 22(24), URL https://www.mdpi.com/1424-8220/22/24/9825

Alyami S, Alharbi R, Azzedin F (2022) Fragmentation attacks an d countermeasures on 6lowpan internet of things networks: Survey and simulation. Sen sors 22(24), URL https://www.mdpi.com/1424-8220/22/24/9825

work page 2022

[5] [6]

Bao F, Chen IR (2012) Dynamic trust management for internet o f things applications. In: Proceedings of the 2012 International Workshop on Self-Awa re Internet of Things, Association for Computing Machinery, New York, NY, USA, Self-IoT ’12, p 1–6, DOI 10.1145/2378023.2378025, URL https://doi.org/10.1145/2378023.2378025 22 Somayeh Sobati-Moghadam

work page arXiv 2012

[6] [7]

Wireless Personal Communications 113:1–17

Elgendy H, Aly SM, Nabil A (2020) Trust-based model for secure routing in wireless sensor networks. Wireless Personal Communications 113:1–17

work page 2020

[7] [8]

Future Internet 16(10), DOI 10.3390/ﬁ16100368, URL https://www.mdpi.com/1999-5903/16/10/368

Fatima M, Rehman O, Rahman IMH, Ajmal A, Park SJ (2024) Towar ds ensemble feature se- lection for lightweight intrusion detection in resource-constrained iot devices. Future Internet 16(10), DOI 10.3390/ﬁ16100368, URL https://www.mdpi.com/1999-5903/16/10/368

work page 2024

[8] [9]

Journal of Systems Architecture

Firoozi F, et al (2021) A trust-based method for secure comm unication in internet of things using edge and fog computing. Journal of Systems Architecture

work page 2021