Darlan Noetzold

This study presents OntOraculum, an ontology designed to organize and analyze performance metrics in smart environments. It categorizes metrics into Hardware, Network, Software, Energy, and SecurityAndReliability domains, defining properties such as dependencies, impacts, and trends to support predictive analysis and dynamic adjustments. Using Protégé for modeling, Semantic Web Rule Language (SWRL) rules facilitate automated inferences, while SPARQL Protocol and Resource Description Framework (RDF) Query Language (SPARQL) queries enable data retrieval from the ontology. Through validation queries, the ontology addresses questions related to network latency, energy efficiency, security, quality of service, and anomaly detection. By unifying performance metrics across diverse domains, OntOraculum enables comprehensive analysis of dependencies and interactions among metrics. The results demonstrate its practical applicability for monitoring, optimizing, and adapting systems in distributed environments.

This article presents the design and deployment of ClimaBogotá v1.2, a climate prediction system tailored for high-altitude urban micro-zones in Bogotá, Colombia. The system combines low-cost IoT sensing, machine learning modeling, and cloud-based orchestration to enable scalable and affordable meteorological forecasting. Its architecture comprises Raspberry Pi-based weather stations, a Random Forest model trained on engineered temporal features, and an n8n-driven automation pipeline for real-time inference and dissemination via Telegram, PostgreSQL, and Grafana. With a Mean Absolute Error of 2.59°C and an R² of 0.6286 on a 30-minute forecast horizon, the system demonstrates both predictive reliability and operational feasibility using free-tier cloud resources. Unlike traditional weather systems, ClimaBogotá emphasizes modularity, adaptability, and cost-efficiency, offering a replicable framework for decentralized climate monitoring in data-scarce urban environments. Temporal misalignment between sensor nodes was identified as the primary constraint, informing future enhancements toward distributed learning strategies.

The rapid expansion of the Internet of Things (IoT) has led to an exponential increase in sensor-generated data, creating challenges for efficient data management and transmission. To address these challenges, SHiELD offers a comprehensive sensor data simulation platform that leverages heuristic techniques such as aggregation, compression, and filtering to streamline data flow without compromising data fidelity. The platform incorporates a suite of advanced predictive models—including ARIMA, LSTM, and Transformer architectures, to accurately forecast sensor behavior and trends. Additionally, SHiELD features fault injection capabilities to evaluate system robustness under adverse conditions. It produces detailed reliability assessments based on metrics evaluating time-series similarity, recovery performance, and transmission quality. Validation experiments, including real-world data acquisition using Arduino-based sensor interfaces and processing on embedded and server platforms, demonstrate that SHiELD’s heuristics can reduce data volume by 8.3% to 13.5% (averaging 9.4%) and lower packet transmission counts by as much as 82.5%. The predictive models integrated within the system achieve strong performance, with F1-scores reaching up to 0.93 and ROC AUC values up to 0.97 for top-performing architectures such as the Transformer and Prophet. Overall, SHiELD serves as an integrated framework for simulating, predicting, and assessing the reliability of IoT sensor data streams.

Self-adaptive systems are increasingly used for effective operation of smart environments, enabling dynamic responses to changing conditions, user needs, and system requirements. This article provides a structured overview of the main strategies for self-adaptation, relevant performance metrics, and key implementation challenges in smart environments.

The emergence of quantum computing challenges traditional security mechanisms, particularly in heterogeneous and resource-constrained IoT and smart environments that must satisfy DN25 requirements. This work introduces a reinforcement learning-driven model for the adaptive selection and orchestration of cryptographic algorithms. Acting as an intelligent decision layer, the system observes contextual, network, and operational metrics to recommend or enforce configurations combining classical schemes, post-quantum cryptography, and Quantum Key Distribution when available. The selection problem is formulated as a Markov Decision Process with state and action spaces aligned with protocol control flows and is embedded into a security middleware with negotiation and fallback mechanisms to ensure interoperability and policy compliance without modifying application logic. Experimental results demonstrate that the model dynamically adjusts key lengths, algorithm families, and security policies according to risk and resource conditions, increasing post-quantum cryptography and Quantum Key Distribution usage by up to 33.4% and 23.9% in high-risk scenarios while favoring low-latency classical or hybrid options for less critical traffic. The system achieves success rates above 78% while maintaining stable latency and resource usage during extended operation.

This paper introduces PRISEC III, a dynamic cryptographic framework designed to balance security and performance in heterogeneous IoT environments. Unlike static, one-size-fits-all approaches, PRISEC III employs a role-based multi-level model that adapts cryptographic strategies according to data sensitivity, device constraints, and network conditions. The framework integrates lightweight symmetric primitives for efficiency, robust asymmetric methods for secure key exchange, and hybrid schemes that combine multiple layers of protection. To enhance long-term resilience, PRISEC III also incorporates post-quantum options, ensuring preparedness against emerging cryptographic threats. Evaluation results demonstrate that the approach achieves scalable security while maintaining feasible computational costs, making it suitable for large-scale and resource-constrained IoT deployments.

Memory optimization in Java applications is essential for performance and scalability. This paper investigates the efficiency of the StringDeduplication parameter in JVM versions 11, 17, and 21, using a Web Crawler developed in Spring Boot. The results show that the efficiency of StringDeduplication decreased from 34.3% deduplication in version 11 to 3.4% in version 21, with an increase in deduplication time from 1,264 ms to 3,439 ms. To mitigate this problem, a custom solution in C was developed for JVM version 21, which increased deduplication to 31.1% and saved 110.2 MB of memory. The main scientific contribution of this work is the identification of the loss of efficiency of StringDeduplication in the latest JVM versions and the proposal of a custom solution that improves string deduplication, offering a viable alternative for developers and software engineers.