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

Somayeh Sobati-Moghadam · 2025

Read Predictive-CSM: Lightweight Fragment Security for 6LoWPAN IoT Networks by Somayeh Sobati-Moghadam (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-01767.txt.

Introduction

Sobati-Moghadam proposes in twenty twenty-five a light defence for split-packet attacks in low-power Internet of Things networks. In these networks, packets are routinely cut into fragments to fit tiny radio frames. Receivers store fragments before checking them, so attackers can flood memory with fake early fragments at little cost.

Her system scores each node's behaviour over time, rewarding steady success and punishing suspect patterns fast. A chained hash checks fragment integrity early, catching tampered sequences before they waste memory. Simulations of injection, replay, and flooding attacks kept delivery rates and energy use steady. The author pitches behaviour-based defence over heavy cryptography for tiny devices.

Catalog record

CREATORSomayeh Sobati-Moghadam
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE66 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-01767

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Predictive-CSM: Lightweight Fragment Security for 6LoWPAN IoT Networks Somayeh Sobati-Moghadam 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 devices 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. 1 Introduction These days, low-power wireless networks are doing a lot of heavy lifting in the world of IoT. Whether it’s managing irrigation systems in agriculture, monitoring machinery in factories, or helping homes run more efficiently, these networks are everywhere. One technology that’s made this possible is 6LoWPAN. It allows tiny devices with minimal memory and power to speak IPv6—essentially giving them a seat at the table on the global internet, even if they’re running on coin cell batteries and a few kilobytes of RAM. But as useful as 6LoWPAN is, it comes with a trade-off. Since these small devices can’t send big packets all at once, the data has to be broken into fragments. That sounds fine in theory, but in practice it creates a weak spot. Devices tend to accept these fragments as they arrive and hold them in memory, even before knowing if the whole message makes sense or comes from someone trustworthy. That small gap—between accepting and verifying—gives attackers just enough room to cause trouble. Since fragments are typically accepted and held in memory before the complete packet is reassembled and validated, attackers can exploit this gap. Even without cryptographic keys or full control of the network, a malicious node can cause serious disruption simply by manipulating how fragments are handled.This opens the door to a range of attacks that target the reassembly buffers of constrained devices. These include not only traditional buffer-reservation attacks but also more nuanced behaviors like clone flooding, header replay, and timing-offset injections—many of which have been observed in recent threat modeling studies Sasi et al (2024) ; Ghubaish et al (2021) . For example, in a simple buffer-reservation attack, an adversary sends incomplete packet fragments (often just the initial FRAG1, which refers to the first fragment of a 6LoWPAN packet that carries essential header information and initiates the reassembly process) slightly before a legitimate transmission. Since devices have very limited buffer space—typically only one or two slots—this alone is enough to block legitimate communication. More advanced strategies involve injecting full but meaningless fragment sequences to exhaust reassembly logic, replaying cloned headers to impersonate trusted nodes, or introducing fragment delays to confuse packet sequencing and induce timeouts. These low-cost, high-impact attacks can silently degrade network performance or cause complete denial of service. Existing mitigation strategies offer only partial solutions. Stateless 6LoWPAN stacks, such as those implemented in Contiki-NG, have no native fragment validation and accept traffic purely based on header structure and timing. Trust-based systems like Chained Secure Mode (CSM) Raoof et al (2020) attempt to mitigate this by filtering packets from untrusted routes, but they operate only at the routing layer and are blind to fragment-level anomalies. On the other end of the spectrum, cryptographic approaches such as SecuPAN Hossain et al (2018a) enforce MAC-based validation of every fragment, but at a cost that is often too heavy for battery-powered or low-RAM devices. In this paper, we present Predictive-CSM , a practical and resource-aware defense framework for 6LoWPAN fragmentation attacks. The system combines two lightweight mechanisms: an adaptive trust model that continuously learns from a neighbor’s fragment-level behavior, and a chained hash validator that verifies the integrity of each fragment incrementally. Unlike traditional models, Predictive-CSM operates directly at the adaptation layer, where fragmentation occurs, enabling the system to identify and discard suspicious fragments before they trigger memory exhaustion or communication breakdowns. We evaluate Predictive-CSM across five carefully designed attack scenarios: early FRAG1 injection (buffer-reservation), full-fragment flooding, header-replay cloning, burst injection (high-rate FRAG1 spam), and late-phase injection (timing-offset spoofing). These scenarios are chosen to reflect both known attacks and realistic adversarial behavior patterns. Simulation results showed that our approach maintains high packet delivery success, rapid adversary detection, and low energy overhead under all attack conditions. By aligning security enforcement with the actual layer where fragmentation occurs, Predictive-CSM closes a longstanding gap in 6LoWPAN security. It offers a deployable, efficient, and adaptive solution that directly addresses the operational constraints of real-world IoT networks while resisting some of the most effective low-layer threats known to date. Paper Organization. The remainder of this paper is structured as follows: Section 2 reviews existing approaches to 6LoWPAN fragmentation security and highlights their limitations. Section 3 introduces the architecture of the proposed Predictive-CSM framework, including the Predictive Trust Engine (PTE) and Fragment Signature Validator (FSV). Section 4 presents the detailed protocol design, including message formats, trust evaluation logic, and fragment validation workflow. Section 5 describes the simulation environment, attack scenarios, and evaluation metrics used to assess performance. Section 6 provides a comprehensive analysis of the experimental results under multiple adversarial conditions. Section 7 develops an analytical model of trust dynamics, buffer utilization, and cryptographic validation. Finally, Section 8 concludes the paper and outlines directions for future work. 2 Related Work Recent advances in securing 6LoWPAN networks have led to various approaches addressing fragmentation vulnerabilities. For instance, Sharma et al. Sharma and Gupta (2021) proposed an adaptive trust management system that leverages machine learning techniques to identify and mitigate attacks on IoT devices, enhancing the resilience of low-power networks. While the machine learning-based trust management system adapts to new threats, it may require extensive training data to perform effectively. In dynamic IoT environments, obtaining sufficient labeled data can be challenging. The machine learning models may introduce computational overhead that is not suitable for resource-constrained devices commonly found in 6LoWPAN networks. The reliance on statistical methods may lead to false positives, causing legitimate traffic to be incorrectly classified as malicious. Li et al. Li and Wang (2022) introduced a lightweight cryptographic protocol specifically designed for 6LoWPAN, which utilizes elliptic curve cryptography to provide secure communication while minimizing computational overhead. Their approach emphasizes energy efficiency, crucial for battery-operated devices. Although their cryptographic protocol is lightweight, the management of elliptic curve keys is still be complex in large-scale IoT deployments, particularly when devices are mobile or frequently join/leave

Read the complete text