This paper article aims to explore the study of different network topologies using Cisco Packet Tracer as the primary investigative tool.

Indexing is one of the most important mechanisms for improving query performance in database systems. Without a proper index, even simple queries can become extremely slow when dealing with large amounts of data. This paper reviews the main indexing techniques used in modern databases from classical tree-based and hash-based structures to bitmap indexes and distributed approaches suited for big data environments. For each technique, the underlying principles, performance characteristics, and limitations are analysed. The paper also includes a comparison table and a discussion of recent trends, such as learned index structures and AI-assisted query optimization. The main takeaway is straightforward: there is no universally best index the right choice always depends on the workload, the data, and the system architecture.

The Spanning Tree Protocol (STP), standardized in IEEE 802.1D, has served as the foundational loop-prevention mechanism in bridged Ethernet networks for over three decades. Despite its widespread adoption, STP exhibits significant limitations with respect to convergence time, bandwidth utilization, and scalability that are increasingly incompatible with the demands of modern enterprise and data center networks. This paper presents a structured comparative analysis of five major Layer 2 protocols in the STP lineage: STP (IEEE 802.1D), Rapid Spanning Tree Protocol (RSTP, IEEE 802.1w), Multiple Spanning Tree Protocol (MSTP, IEEE 802.1s), Shortest Path Bridging (SPB, IEEE 802.1aq), and Transparent Interconnection of Lots of Links (TRILL, RFC 6325). The analysis is conducted along four primary dimensions: convergence time, scalability, operational complexity, and security considerations. Drawing exclusively on peer-reviewed literature, IETF/IEEE standards, and published operational benchmarks, the paper concludes with a structured evaluation of the relevance of these protocols in Software-Defined Networking (SDN) environments and modern Ethernet fabric architectures, proposing a conceptual decision framework for protocol selection.

Machine learning-based network intrusion detection systems are often optimized using aggregate accuracy, although operational security depends on the reliable detection of rare, high-impact attacks. This paper proposes a data-preserving intrusion detection framework that combines hybrid feature selection, heterogeneous ensemble learning and cost-sensitive optimization for imbalanced multi-class attack detection. The method first applies Mutual Information filtering and Recursive Feature Elimination to reduce the NSL-KDD feature space from 122 one-hot encoded attributes to 25 discriminative features. Four classifiers, Random Forest, XGBoost, Support Vector Machine and K-Nearest Neighbors, are evaluated individually, and a stacking ensemble is constructed using Logistic Regression as a meta-learner. Class imbalance is addressed by balanced class weighting rather than by synthetic oversampling, preserving the original minority-class observations. Experiments on the NSL-KDD benchmark show that the proposed cost-sensitive configuration improves rare attack recognition, most notably increasing U2R recall from 0.00% to 35.82% (24 of 67 test instances) for the stacking ensemble; this improvement, together with the accompanying weighted F1-score change from 0.7120 to 0.7214, is statistically significant under the Wilcoxon signed-rank test across repeated random seeds, and both values are reported with their variability rather than as single point estimates. SVM obtains the largest global gain, with a 7.06 percentage point improvement in weighted F1-score. The results show that cost-sensitive learning is a simple and practical mechanism for improving rare-attack visibility, but also reveal a remaining limitation for R2L detection, where feature overlap with Normal traffic remains substantial. The revised validation design explicitly includes direct resampling baselines, repeated-seed evaluation, statistical significance testing, feature-subset sensitivity analysis, and absolute true-positive counts for R2L and U2R in order to avoid overinterpreting marginal point-estimate gains. All experiments, including the resampling comparison, the component ablation, the feature-subset sensitivity analysis and the repeated-seed statistical evaluation, are executed on the complete KDDTrain+ training set of 125,973 instances under a single unified protocol, so that every reported per-class value refers to the same experimental setting. The revised study additionally reports probability-level evaluation for the primary model, including class-level PR-AUC, precision-recall curves and a U2R threshold and alert-budget analysis, and validates the framework externally on the UNSW-NB15 benchmark, where balanced class weighting raises the recall of the rarest categories (Worms, Shellcode, Backdoor) from near-zero baseline levels to 69–96% under an identical protocol.

This paper proposes a block-adaptive reversible data hiding scheme based on Prediction Error Expansion (PEE) that eliminates the predictor map overhead entirely. In existing block-adaptive methods, the optimal predictor is selected per block and its index must be transmitted as side information, which reduces net embedding capacity. The proposed method removes this requirement by exploiting the causal raster-order processing: the predictor for each block is chosen by evaluating the embedding capacity of the preceding block, which is in an identical watermarked state for both encoder and decoder. As a result, both parties independently arrive at the same predictor decision without any communication. The first block uses MED by hardcoded convention, also requiring zero transmitted bits. The scheme applies one predictor uniformly to all pixels within each block, including border pixels, with no fallback required. Experiments on three standard benchmarks — Kodak PhotoCD (24 images), USC-SIPI (39 images), and BSD68 (68 images) — show that the proposed CausalBlock-16 consistently achieves 91–95% of the best fixed predictor (Linear) at threshold T=3, while matching its overhead to that of fixed-predictor baselines. On texture-rich images CausalBlock outperforms Linear, and on individual smooth images it reaches up to 99% of Linear capacity. Full reversibility is formally proven and experimentally confirmed on all 131 benchmark images across all tested configurations.

The mandatory introduction of Romania’s national e-invoicing system, ANAF e-Factura, in January 2024 has reshaped B2B document exchange in the retail sector, but suppliers still operate in parallel with two proprietary electronic data interchange (EDI) platforms, EDINET and DocProcess, which increases integration complexity. This paper presents the architecture, implementation, and evaluation of a custom Laravel-based B2B platform developed to automate commercial workflows across these three channels. The system supports XML purchase order ingestion and normalization, product identifier resolution, unified order persistence, platform-specific invoice XML generation, and ANAF SPV submission via SmartBill and Oblio REST APIs. A comparative analysis of real production XML documents showed full field-level overlap across 21 invoice data dimensions, with the main differences between systems related to entity identification schemes rather than business information content. During 2025, the platform processed 1247 EDI purchase orders and achieved an 87.30% fully automated processing rate, reaching 94.60% by year-end through progressive product catalog enrichment. The results indicate that ANAF e-Factura is technically capable of covering the core invoice exchange function currently duplicated by proprietary EDI platforms, while their coexistence continues to impose additional integration effort and slows SME digital transformation, particularly for small and medium-sized suppliers.