Jong-Hyouk Lee

The revised EU Data Act requires the incorporation of smart contract kill switch mechanisms. Nevertheless, most existing designs are either confined to single chain architectures or dependent on centralized control, thereby conflicting with the decentralization principles of blockchain systems and limiting their applicability in cross-chain environments. This limitation becomes particularly critical in IoT-driven industrial contexts, where security incidents demand synchronized execution control across multiple chains. Under such circumstances, single chain oriented kill switch mechanisms inherently suffer from structural constraints. To address these challenges, we propose an integrated Layer 0 and Layer 1 smart contract kill switch framework. The proposed framework enables distributed decision-making across heterogeneous blockchain systems by supporting diverse consensus algorithms, including PoW, PoS, PBFT, Paxos, and Raft. Cross-chain propagation of kill switch activation signals is securely facilitated through interoperability solutions such as CCMP and LayerZero. Importantly, activation decisions made on one chain are not mandatorily enforced on others; instead, each chain independently determines whether to execute the kill switch via its own consensus process, thereby preserving both autonomy and decentralization. For performance evaluation, repeated experiments were conducted across 50 use cases derived from exhaustive combinations of two interoperability mechanisms and five consensus algorithms. The analysis considered average activation time, gas consumption, and message delay, alongside Time to Activation (TTA) and tail-latency metrics to assess real-time responsiveness. Network experiments using Linux tc-netem demonstrate that even under severely degraded network conditions, the cross-chain p99 TTA is bounded at 12.2596 s, establishing a predictable worst-case upper limit. These results highlight the structural robustness of the proposed framework by guaranteeing a bounded reaction time even in environments characterized by increased latency and consensus overhead, rather than merely improving average-case performance.

Recent warfare is transitioning from traditional domains like land, sea, and air to cyber warfare. This transition suggests that efforts will be made to attack and neutralize a nation's critical information infrastructure, and steal sensitive data stored therein to secure a strategic advantage. Recent instances have shown attackers bypassing air-gap technologies-which physically or logically disconnect networks from the Internet-by setting up covert channels to illicitly transfer sensitive data via unconventional routes. This challenges the prevailing confidence in current network separation technologies, a concern that the industrial sector has yet to fully appreciate. This paper delves into previously researched acoustic signal-based air-gap attack technologies. It identifies their limitations and introduces an advanced acoustic signal air-gap attack designed to minimize error rates even in noisy environments where signal distortion is caused by environmental interference, such as ambient noise. This is achieved by optimizing the exploration for distorted frequency bands, thus highlighting the potential impact of air-gap attack technology. The proposed technology gradually increases the exploration range from 1 Hz, using the frequency utilized for data exfiltration to identify the optimal band. This approach allows for precise data leakage through targeted signal processing, even in environments with significant interference. Experimental results demonstrate that this advanced approach, which focuses on identifying the optimal frequency band, maintains a Bit Error Rate (BER) of no more than 1.5625% up to a distance of 40m, showing a 7.53% improvement in time efficiency over traditional band-pass filter-based methods. As air-gap bypassing inherently limits network bandwidth and transmission speed compared to standard TCP/IP communications, having time-efficient air-gap attacks offers practical benefits and enhances the likelihood of evading detection. Additionally, after applying a formula to derive the trend of sound pressure attenuation with distance, the paper compares these theoretical predictions with actual measurements, confirming that the proposed attack reaches the maximum potential distance for air-gap attacks. This demonstrates that the proposed attack has the highest possible performance in relation to distance.

Recently, advanced closed-circuit televisions have been widely installed for security and crime prevention purposes, leading to the large-scale adoption of facial recognition systems. However, the indiscriminate use of such technologies has resulted in excessive intrusion into individuals’ private lives, prompting ongoing discussions around privacy infringement. In response, Homomorphic Encryption (HE)-based facial recognition technologies—which aim to ensure both privacy and security—have been actively researched and developed. Nevertheless, existing researches have limitations in that they did not seriously consider the possibility of original image leakage and failed to incorporate color information into the similarity measurement process. To address these issues, we propose a mechanism that considers not only facial landmarks but also color information in facial recognition, applying HE in a way that causes actual image modification in the facial area. This ensures that privacy can still be protected even in the event of facial data leakage. In the proposed mechanism, similarity values are independently derived from both facial landmarks and aligned color information. After decrypting each homomorphically encrypted value, a final Dual-Layer similarity is computed by combining both, enabling the identification of the facial image with the highest similarity. To achieve highly accurate and reliable final similarity computation, we propose a Dual-Layer similarity derivation formula based on the Softmax function, which is widely used as a core element in deep learning for multi-class classification. The effectiveness and necessity of the proposed method are demonstrated through comparisons with the arithmetic mean, weighted arithmetic mean, and harmonic mean. Furthermore, since most existing HE schemes are limited to supporting addition and multiplication operations, we analyze similarity derivation algorithms that comply with these constraints. Through a comparative analysis based on computational efficiency, we demonstrate that Euclidean Distance is well-suited for efficient HE-based facial recognition.