Category Archives: Uncategorized

Authors: Artur Eduardovich Karapetyan, Lusine Rafikovna Minasyan, Hovhannes Grigorievich Manukyan, Ani Serobovna Avetisyan, Vardan Levonovich Sahakyan

Abstract: The exponential growth of genomic data presents immense challenges in terms of storage, processing, and analytics. Hybrid cloud systems—combining on-premises resources with scalable cloud services—offer a compelling solution for addressing these computational demands. This paper presents a unified architectural model designed to optimize genomic data analytics in hybrid cloud environments. By integrating containerized bioinformatics workflows, secure data orchestration mechanisms, and AI-driven scheduling, the proposed framework ensures agility, scalability, and compliance. We explore the role of cloud bursting for peak genomic analysis workloads, address data residency and regulatory concerns, and demonstrate performance improvements across typical use cases such as variant calling and gene expression analysis. This architecture supports real-time analytics, secure collaboration, and cross-institutional data sharing in the genomics domain.

DOI: https://doi.org/10.5281/zenodo.16354292

Published by: vikaspatanker

Authors: Zamira Sadullaevna Rajabova, Otabek Abduvohidovich Madrahimov, Dilshod Jamolovich Saidov, Malika Rasulovna Kadirova

Abstract: The evolution of biomedical data analytics has been closely tied to the scalability and reliability of server infrastructure. Logical Domains (LDOMs), a virtualization technology native to Oracle Solaris, offer hardware-level isolation and performance efficiency that align well with the computational demands of machine learning (ML) in biomedical applications. This research investigates the deployment, optimization, and execution of various ML models within LDOM-enhanced server environments specifically tailored for high-throughput biomedical workloads. It evaluates the architectural benefits, virtualization overhead, and performance stability when applying ML algorithms for genomics, diagnostics, and health informatics. The findings suggest that LDOM-based infrastructures not only support secure multitenancy for ML pipelines but also enable tunable resource allocation strategies for precision performance in real-time medical contexts.

DOI: https://doi.org/10.5281/zenodo.16352646

Published by: vikaspatanker

Authors: Oleh Mykhailovych Hrytsenko, Iryna Volodymyrivna Lysenko, Denys Ivanovych Sydorenko, Viktoriia Andriivna Kravets

Abstract: The integration of Edge-AI with wearable biomedical devices like Myobioscan is reshaping the landscape of real-time clinical diagnostics and patient monitoring. This paper explores how embedding artificial intelligence at the device edge enables low-latency, high-frequency processing of biosignals such as electromyography (EMG), electrocardiography (ECG), and motion patterns. The combination of Myobioscan’s compact sensor technology with on-device AI accelerators facilitates proactive health assessments, early anomaly detection, and decentralized clinical interventions. By reviewing recent deployments and experimental models, this study identifies key performance metrics, data handling architectures, and regulatory considerations in deploying Edge-AI for mobile health. The findings point toward a scalable and responsive healthcare ecosystem driven by distributed intelligence at the physiological interface.

DOI: https://doi.org/10.5281/zenodo.16352134

Published by: vikaspatanker

Authors: Andrei Nikolayevich Petrovski, Ekaterina Leonidovna Sokolova, Vladislav Dmitrievich Morozov, Irina Sergeyevna Volkova

Abstract: Cardiovascular diseases (CVDs) remain the leading cause of mortality worldwide, necessitating prompt and accurate risk detection for timely intervention. This research presents a scalable artificial intelligence (AI) infrastructure designed to support real-time cardiovascular risk detection using streaming medical data. The proposed architecture integrates distributed data ingestion, edge AI processing, and cloud-based model orchestration to ensure both low-latency diagnostics and high system reliability. Using a combination of convolutional neural networks (CNNs) for ECG signal analysis and gradient-boosted trees for patient history correlation, the system demonstrates improved predictive accuracy. Performance benchmarks show efficient scaling across multiple nodes, enabling high-throughput analysis essential for deployment in emergency and critical care settings. The paper evaluates model deployment on Kubernetes, real-time data flow with Apache Kafka, and compliance with healthcare data privacy regulations. The study concludes with recommendations for integrating this AI infrastructure into hospital networks and telemedicine platforms.

DOI: https://doi.org/10.5281/zenodo.16317726

Published by: vikaspatanker

Authors: Andrei Nikolayevich Petrovski, Ekaterina Leonidovna Sokolova, Vladislav Dmitrievich Morozov, Irina Sergeyevna Volkova

Abstract: Cardiovascular diseases (CVDs) remain the leading cause of mortality worldwide, necessitating prompt and accurate risk detection for timely intervention. This research presents a scalable artificial intelligence (AI) infrastructure designed to support real-time cardiovascular risk detection using streaming medical data. The proposed architecture integrates distributed data ingestion, edge AI processing, and cloud-based model orchestration to ensure both low-latency diagnostics and high system reliability. Using a combination of convolutional neural networks (CNNs) for ECG signal analysis and gradient-boosted trees for patient history correlation, the system demonstrates improved predictive accuracy. Performance benchmarks show efficient scaling across multiple nodes, enabling high-throughput analysis essential for deployment in emergency and critical care settings. The paper evaluates model deployment on Kubernetes, real-time data flow with Apache Kafka, and compliance with healthcare data privacy regulations. The study concludes with recommendations for integrating this AI infrastructure into hospital networks and telemedicine platforms.

Published by: vikaspatanker

Authors: Guram Shalvovich Danelia, Nino Giorgievna Kalandadze, Levan Besarionovich Mchedlidze, Salome Iraklievna Tsereteli

Abstract: The convergence of artificial intelligence (AI) and health informatics has the potential to revolutionize clinical decision-making, disease surveillance, and personalized medicine. This study explores the integration of AI workflows with existing health informatics pipelines, examining both the transformative opportunities and the critical challenges associated with such integration. By analyzing case studies from electronic health record (EHR) systems, bioinformatics pipelines, and radiological imaging networks, we identify architectural patterns that enable seamless AI integration. Additionally, the research addresses the barriers posed by data heterogeneity, workflow fragmentation, regulatory compliance, and algorithm interpretability. The findings suggest that while AI offers immense benefits in improving healthcare outcomes and operational efficiency, a strategic, interoperable, and ethically grounded approach is necessary for scalable implementation in health informatics infrastructures.

DOI: https://doi.org/10.5281/zenodo.16315871

Published by: vikaspatanker

Authors: Sergey Artyomovich Mamedov, Yelena Ramizovna Isayeva, Anar Fikret oglu Mahmudov, Kamilla Rauf qizi Veliyeva

Abstract: Biomedical data centers serve as the backbone of modern healthcare analytics, precision medicine, and hospital informatics. As the volume of healthcare data surges, the need for scalable, secure, and efficient computing infrastructure becomes paramount. Server virtualization has emerged as a critical enabler in this space, offering resource abstraction, fault tolerance, and operational flexibility. This study performs a comparative assessment of leading server virtualization techniques—namely hypervisor-based (e.g., KVM, VMware ESXi), container-based (e.g., Docker, LXC), and hybrid models—based on key parameters such as performance, scalability, resource utilization, latency, and compliance with biomedical data handling norms. Benchmarks using real-world datasets, including EHRs and PACS workloads, reveal that no single approach dominates across all metrics, emphasizing the need for context-driven infrastructure design.

DOI: https://doi.org/10.5281/zenodo.16314536

Published by: vikaspatanker

Authors: Rahmatulloh Tohirovich Saidov, Mavjuda Gafurovna Khudoerzoda, Akmal Ziyodulloevich Kholov, Shahrukh Nasrulloevich Nazarov, Dilnoza Mahmadaliyevna Qalandarova

Abstract: As genomic data repositories expand rapidly with the growing need for precision medicine and population-scale genomics, managing the integrity, security, and regulatory compliance of these repositories has become a paramount concern. This paper presents a compliance-centric approach to automating server infrastructure specifically tailored for genomic data management. We explore the integration of Unix-based server automation tools with security-first policies and standards such as HIPAA, GDPR, and ISO 27001. The study outlines how configuration management, access control automation, logging, and continuous compliance auditing are implemented to ensure operational resilience and regulatory alignment. By leveraging scripting, cron-based scheduling, and policy-as-code frameworks, genomic data infrastructures can be both scalable and secure. The proposed automation model reduces human error, enhances traceability, and allows for real-time response to compliance deviations, making it a critical foundation for modern biomedical computing environments.

DOI: https://doi.org/10.5281/zenodo.16313840

Published by: vikaspatanker

Authors: Elen Rafayelovna Sargsyan, Hayk Vahagnovich Ghazaryan,, Anzhela Viktorovna Grigoryan, Karen Samvelovich Melikyan, Tatevik Aramovna Harutyunyan

Abstract: The explosive growth of genomic and proteomic datasets has propelled bioinformatics into the enterprise computing domain, demanding scalable, secure, and high-performance infrastructure. Traditional physical server models have proven inadequate for managing the dynamic and compute-intensive nature of bioinformatics workflows. In response, AI-powered virtualization models are emerging as transformative solutions, combining intelligent workload orchestration with flexible virtual environments. This paper investigates how artificial intelligence enhances virtualization strategies in enterprise bioinformatics settings by enabling predictive resource allocation, automated fault detection, and real-time optimization. Through architectural analysis and case study evaluation, the research presents a practical framework for deploying AI-integrated virtual infrastructure that meets the evolving needs of large-scale biological computation.

DOI: https://doi.org/10.5281/zenodo.16312734

Published by: vikaspatanker

Authors: Natalia Ivanovna Baranova, Dmitry Alekseevich Tikhonov, Yulia Sergeevna Pankratova, Ivan Mikhailovich Rogozin

Abstract: Electronic Health Records (EHRs) are foundational to modern healthcare systems, but they are also lucrative targets for cyberattacks due to the sensitivity of medical data. Ensuring the confidentiality, integrity, and availability of EHRs requires robust server-level defenses. This study investigates the implementation of adaptive Unix server hardening models tailored for healthcare environments. It outlines a layered approach to security that integrates dynamic configuration baselines, continuous monitoring, and compliance mapping to standards like HIPAA, HITRUST, and NIST. Through adaptive hardening strategies, including automated shell scripts, auditing frameworks, and anomaly detection, we propose a defense-in-depth model that significantly enhances EHR security posture. Real-world use cases and benchmarks validate its practicality in live healthcare infrastructures.

DOI: https://doi.org/10.5281/zenodo.16312367

Published by: vikaspatanker