Research investigated behavior-based multi-class malware family classification using the large-scale BCCC-Mal-NetMem-2025 dataset. Using a MSFormer (Multi-Sequence Fusion Classifier), transformer-inspired architecture designed for multi-class malware family classification through an attention mechanism, harnessing the complementary strengths of network and memory modalities for more accurate and robust family-level classification. Our fusion framework integrates multi-head attention mechanisms with Word2Vec embeddings and applies SMOTE to address class imbalance, followed by a fusion layer for improved classification, unlike previous work that relied solely on a single-source dataset...

Network security through cryptographic algorithm and protocols remains the best suitable method for maintaining confidentiality such that in this paper different cryptographic algorithms are compared, analysis is performed and results are obtained for further operations on the basis of outcomes. Network Security; Cryptography; RSA; DES; AES; Blowfish; 3-DES