distributed systems

Deploying small language models (e.g., SLMs) on edge devices has become increasingly viable due to advancements in model compression and efficient inference frameworks. Running small models offers significant benefits, including privacy through on-device processing, reduced latency, and increased autonomy. This paper conducts a comparative review and analysis of Node.js inference frameworks that operate on-device. It evaluates frameworks in terms of performance, memory consumption, isolation, and deployability.

The article discusses modern information security technologies in distributed and asymmetric systems, as well as problems arising from their implementation in the context of growing cyber threats. An analysis of cryptographic methods, authentication systems, access control, and intrusion detection has been provided. Particular attention has been paid to the limitations of existing technologies and promising areas for their improvement, in particular the use of machine learning methods, block chain technologies, and the Zero Trust concept.

The article proposes a comprehensive approach to solving the data protection problem in decentralized distributed information storage systems based on blockchain technology. A conceptual ‘SecureChain’ model has been developed that integrates modern cryptographic protection methods with programmable smart contract logic for automated access management and data integrity assurance. The model employs a multi-level architecture including data layer, smart contract layer, network interaction layer, and user interface.

The development of smart contracts in blockchain networks has enabled the creation of sophisticated decentralized finance (DeFi) protocols, encompassing decentralized exchanges, lending platforms, and algorithmic crypto-assets. Despite decentralization and transparency, blockchain networks do not guarantee a predictable transaction execution order, leading to the emergence of the phenomenon known as Maximal Extractable Value (MEV) – an additional profit extracted by certain network participants who influence transaction ordering.

This paper substantiates an approach to performing calibration procedures for measurement channels on-site in distributed cyber-physical systems. A conceptual model of a local metrological supervision system is proposed, which provides calibration considering real operating conditions. Special attention is paid to the use of code-controlled measures to generate reference signals and a centralized module for remote control, data processing, and storage.

The rapid expansion of the Internet of Things (IoT) has resulted in a substantial increase of diverse data from distributed devices. This extensive data stream makes it increasingly important to implement robust and efficient real-time anomaly detection techniques that can promptly alert about issues before they could escalate into critical system failures.

The increasing complexity of cyber threats requires the development of effective methods for detecting and classifying attacks in network traffic. This study analyzes the effectiveness of three popular machine learning algorithms: Random Forest, which is used for anomaly detection, Support Vector Machines (SVM), which performs cyber threat classification, and autoencoders, which are used for data preprocessing and deep traffic analysis.

The article discusses key information security principles, focusing on confidentiality, integrity, availability, traceability, and the DIE model (Distributed, Immutable, Ephemeral). Confidentiality emphasizes the importance of secrecy and controlling access to prevent sensitive information from misappropriation. Integrity ensures that data remains accurate and trustworthy, with measures to prevent unauthorized modifications.

The article addresses the issue of data protection in information and communication systems amid the growing volume of traffic and the increasing number of cyber threats, necessitating improvements in the effectiveness of intrusion detection and prevention systems. Various types of Intrusion Detection Systems (IDS) and Intrusion Prevention Systems (IPS), their advantages, and disadvantages are considered. The methods of threat detection are analyzed, including signature-based methods, anomaly detection methods, and machine learning-based methods.