Journal of Science and Technology on Information security

Refining Galois/Counter Mode for 256-Bit Blocks: Multilinear Authentication and Security Proofs

Nguyen Tuan Anh — 2026-06-24

This paper investigates the feasibility of extending the authenticated encryption mode GCM to a 256-bit block size to increase the amount of authenticated data and enhance security in large-scale information processing systems. Based on an analysis of the limitations of the standard GCM construction, particularly the restriction on the amount of data processed per invocation and the weaknesses arising from the linearity of the GHASH function, we propose a GCM-256 variant employing a 256-bit block cipher together with a multilinear authentication mechanism, replacing the conventional linear GHASH function.

On the Uniqueness of RSA Keys in EJBCA PKI Systems

Nguyen Le Minh — 2026-06-24

RSA deployments in EJBCA-based public key infrastructures (PKIs) rely on X.509 Subject Key Identifiers (SKIs) to prevent reuse of identical public keys. However, SKI-level checks fail to detect shared-prime collisions, where distinct RSA moduli reuse a prime factor and become factorable via batch GCD analysis. This paper examines why existing SKI mechanisms cannot guarantee RSA key uniqueness and how such weaknesses propagate through enrollment and renewal processes in EJBCA.

We propose two complementary controls that extend uniqueness from the public key to RSA’s secret parameters. The first is a secrecy-preserving duplication check, where one-way commitments to the primes (p, q) are embedded in the certificate request and verified against a commitment registry to block reuse without exposing secret values. The second is an identity-scoped RSA generation algorithm that deterministically maps subject identifiers to disjoint prime-search intervals, ensuring well-spaced and non-overlapping primes across users.

RIS Assisted Anti Jamming Backscatter Communication Using Deep Reinforcement Learning

Le Hoang Hiep — 2026-06-24

Backscatter communication is a promising low-power solution for large scale Internet of Things (IoT) networks; however, it is highly susceptible to intentional jamming and dynamic interference. This paper proposes a Deep Reinforcement Learning (DRL) - based anti-jamming framework for reconfigurable intelligent surface (RIS) - assisted backscatter systems. The problem is formulated as a Markov decision process, where an agent adaptively optimizes RIS reflection coefficients without prior environmental knowledge. A deep Q-network with experience replay and a target network ensures stable learning in high-dimensional state spaces. Simulation results show convergence within 600 – 800 episodes. Compared with random RIS configurations and RIS-free systems, the proposed scheme improves SINR by 3 – 6 dB and over 8 dB, respectively, under moderate-to-strong jamming. Moreover, BER is reduced by nearly one order of magnitude, and throughput increases by 25% -40%. These results demonstrate robust and adaptive interference mitigation for future IoT and 6G networks

DVA-IR: Runtime Integrity Attestation for Secure VNF Operation in NFV

Vu Ngoc Hoa — 2026-06-24

Network Function Virtualization (NFV) improves flexibility and cost efficiency, but it also creates new security risks at the virtualization, VNF, service-chain, and orchestration layers. This paper focuses on a specific NFV security problem: detecting and containing runtime compromise of Virtual Network Functions (VNFs), including unauthorized binary, configuration, or memory-state modification, invalid attestation responses, and no-response events. We propose Dynamic VNF Attestation and Instant Response (DVA-IR), a framework that combines periodic remote attestation of running VNFs with MANO-driven isolation and replacement of compromised VNFs. Unlike access-control-only or signature-based mechanisms, DVA-IR verifies runtime VNF integrity and immediately updates service-chain forwarding behavior when compromise is detected. A Monte Carlo simulation compares DVA-IR with manual monitoring, signature-based IDS/IPS, and DVA-only attestation. The results show that DVA-IR achieves high detection quality, rapid containment, and significantly reduced service downtime, while its centralized orchestration overhead remains manageable for small-to-medium and medium-to-large NFV deployments.

A Novel Model of Comprehensive Data Encryption for the Website Using Blockchain

Pham Van Huong — 2026-06-24

The paper proposes a comprehensive data security model for blockchain-based web applications. This model can be used as a general template for Web3 applications. The model consists of two parts: a blockchain core with integrated database encryption modules, replacing Fabric CA; and an application part that also integrates file encryption, database encryption, and digital signatures. The proposed model was tested on a VBCC management website using Hyperledger Fabric. File and database encryption uses AES, and digital signatures use ECDSA. To improve performance, we also replaced the GolevelDB database management system with LevelDB. Experimental results confirm the accuracy and good performance.

Lightweight Linear Layers with High Branch Number for Ascon-Like Permutations

Hoang Dinh Linh — 2026-06-24

The linear layer is a fundamental building block of substitution-permutation network (SPN) based ciphers, responsible for the diffusion of differences and linear masks across the state. In this paper, we analyze and evaluate several lightweight linear layers through the lens of branch number theory, focusing on constructions that rely solely on XOR operations and cyclic rotations, such as the Column Parity Mixer (CPM), the Twin Column Parity Mixer (TCPM), the Symmetric Twin Column Parity Mixer (STCPM), and the Multiple Rows Mixer (MRM). Based on these analyses, we propose new linear layers that achieve higher branch numbers while requiring fewer XOR operations than existing designs. We then evaluate the proposed constructions when instantiated in Ascon-like permutations operating on a 5 x 64-bit state, comparing Ascon, Gaston, Gaston-S, and Hsilu in terms of trail weight bounds. Our results confirm that the proposed linear layers achieve: a column differential branch number 7 at 2.8 XOR/bit (Construction 1), branch number 7 with improved linear resistance at 3 XOR/bit (Construction 2), and branch number 17 at 4 XOR/bit (Construction 3). In particular, the final structure provides a substantially higher column branch number compared to TCPM and STCPM at a similar XOR implementation cost

An Explainable Hybrid Framework for Detecting Harmful Short Comments on Vietnamese Social Media Using TF-IDF and Lightweight Transformers

Ha Van Muon — 2026-06-24

The proliferation of harmful content on Vietnamese social media platforms necessitates effective detection systems for short, concise, and slang-heavy texts. This study proposes the Explainable Hybrid Framework (EHF) for enhanced detection and interpretability. On an expert-annotated dataset of 15,162 samples, EHF with SVM achieves an F1-Score of 92.5% and Accuracy of 92.6%, outperforming strong baselines such as PhoBERT on our dataset, while state-of-the-art models like ViHateT5 are discussed on external Vietnamese benchmarks. Notably, it reduces false positives by 30% and false negatives by 25% versus traditional methods, with 45ms inference time. Our primary contribution is an efficient, fast, and explainable solution that addresses the critical challenge of automated moderation for short Vietnamese texts.

Impact of Cipher-Core Selection on XTS-based Secure Network-Attached Storage Systems: A Comparison Study of MKV and AES

Trần Văn Khánh — 2026-06-24