мікросервісна архітектура

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.

This article presents a new scalable system for automated modeling of physical processes and experiment management based on microservice architecture and cloud technologies. The proposed platform addresses the growing need for flexible, cost-effective, and highly scalable computational solutions for physical research. The system utilizes Amazon Web Services cloud infrastructure with containerized microservices to provide automated resource allocation, experiment orchestration, and laboratory equipment integration.

The article examines the security of microservice architectures in the context of common vulnerabilities that arise in distributed systems. The authors analyze the essence of the microservice approach, which, despite its advantages in scalability and flexibility, introduces new challenges in the field of cybersecurity. The main focus is on issues of access management, network configuration, and data protection during transmission between services, which can create additional attack vectors.

An approach to counteracting cyberattacks based on state machines within a microservice architecture is suggested. It focuses on intelligent analysis of actual and possible intrusions. The approach is devised for applications with a microservice architecture deployed on the Kubernetes platform. For purposes of the study, a special dataset has been developed. We have reproduced selected common vulnerabilities and exposures reported in 2024 and collected network traffic of intrusion cyberattacks based on them.

This article presents the development of a recommender system for recruiting personnel and vacancies to improve the efficiency of the hiring process. The proposed system has integrated collaborative and hybrid filtering methods to provide personalized job recommendations. Collaborative filtering model has analyzed historical data, identifying patterns by detecting connections between user information and job content.

This study examines the capabilities and limitations of small, 4-bit quantized language models that run locally on Apple Silicon. Four models have been benchmarked on a dataset of natural language prompts, with runtime metrics including inference time, memory usage, and token throughput, as well as output behavior. The study provides an empirical assessment of the feasibility of deploying language models on resource-constrained devices.

Microservices represent a software development approach, a variation of service-oriented architecture, that structures an application as a collection of loosely connected services. The aim of this work is to explore the design and implementation methodology for information systems using micro-interfaces to enhance development quality and speed while simplifying their usage. This work proposes a method for transitioning from a monolithic software architecture to a microservice architecture.

Integrating AI agents within serverless architectures offers a modern approach to deploying and executing intelligent applications. Leveraging the advantages of serverless computing, AI agents can dynamically respond to varying workloads without the overhead of managing the underlying infrastructure. This article explores the concept of scalable serverless AI agents in the cloud, detailing their architecture, benefits and drawbacks, challenges, and real-world applications. The paper provides advantages and drawbacks of the serverless approach.

The emergence of microservice architecture has revolutionized software development practices by decentralizing components, facilitating scalability, and enabling agility in system design and deployment. There are some benefits of incorporating microservices instead of a single server, however, distributed components introduce extra constraints and complexities in maintaining data consistency as well. As microservices interact independently, coordinating data updates across multiple services becomes challenging, particularly in scenarios where transactional integrity is required.