MANAGING DATA TRANSFER COSTS IN HYBRID CLOUD INFRASTRUCTURES: APPLIED RESEARCH

Elastic Data Platform Architecture for Multi-Cloud IT Cost Optimization and Performance

In today’s dynamic business environment, organizations are increasingly relying on multi-cloud strategies to achieve flexibility, cost efficiency, and scalability. However, managing and optimizing IT costs while ensuring optimal performance across multiple cloud environments remains a complex challenge. This paper explores the concept of an Elastic Data Platform (EDP) as a solution for multi-cloud IT cost optimization and performance. By leveraging the inherent elasticity of cloud resources, this architecture provides the ability to scale data infrastructure efficiently while maintaining high performance levels. We discuss the key design principles of an EDP, including data distribution, workload optimization, auto-scaling, and cost analytics, and how these can be implemented across multiple cloud providers. Additionally, we analyze real-world use cases, benefits, and challenges associated with this architecture. This paper aims to provide insights into how businesses can optimize both costs and performance in a multi-cloud environment using an Elastic Data Platform.

Towards Practical Latency SLOs on Cloud Data Warehouses

Enhancing the Kubernetes Scheduler: A State-of-the-Art Review from Cloud to Edge

A comprehensive review of Kubernetes scheduling strategies published between January 2023 and January 2026 is presented and a multi-dimensional taxonomy is established that categorizes scheduling approaches based on common objectives, modification methods, optimization methodologies, targeted workloads, evaluation methods, scheduling scopes, and performance metrics.

Enabling Real-Time Transaction Processing in Distributed Cloud Systems: Architectural, Operational, and Societal Dimensions

Real-time transaction processing has emerged as a foundational requirement of modern digital economies, underpinning payment networks, trading platforms, and fraud-detection systems that collectively handle trillions of dollars in daily transaction volume. As cloud-native architectures displace monolithic legacy systems, organizations face a dual imperative: delivering consistently sub-millisecond latency at global scale while simultaneously reducing the energy footprint of the infrastructure that sustains these workloads. Systems distributed across clouds using the architecture based on microservices, event-driven pipelines, containerization, and intelligent orchestration have been shown to be able to meet both requirements, although the engineering considerations that go into making this possible have not been extensively studied in the literature. This article investigates the architectural approaches, infrastructure optimization methods, and performance techniques that contribute to enabling high throughput and energy efficiency in transactional systems operating in real time. Using empirical benchmarks and advancements made in cloud computing technology and scheduling, the article further analyzes the social consequences of adopting such a system, including financial inclusion, sustainability, and trust.

Cost-Optimized Database Security Architectures for Multi-Cloud Environments: A Unified Pattern and Decision Model for Regulated Workloads

Aim: This study aims to develop a provider-neutral security architecture and a quantitative decision model for evaluating alternative approaches to securing regulated database workloads across multiple public clouds. Methods: The study employs a model-driven comparative evaluation rather than an empirical research design. Three PCI DSS v4.0- and HIPAA-regulated workload profiles are evaluated across centralized, fragmented, and hybrid multi-cloud security architectures. The analysis integrates published cloud-service pricing, task-level estimates of operational effort, a five-theme compliance-coverage rubric, and sensitivity analysis covering variations in cost, labor rates, tooling requirements, log volumes, and data-residency assumptions. All inputs used in the worked example are reproduced in the study to facilitate transparency and replication. Results: The modeled results indicate that the centralized UMDSP architecture costs 20.7%–25.2% less annually than the fragmented architecture, with a 22.1% cost reduction in the worked example, while maintaining comparable overall compliance-control coverage. However, the centralized approach provides greater audit-evidence coverage, reaching 94% compared with 83% for the fragmented architecture. The cost advantage remains relatively stable under variations in log volume and labor rates but declines to approximately 12% - 18% under conservative assumptions concerning operational effort and tooling. The advantage also decreases when data-residency requirements necessitate duplicated regional analytics. The findings are based entirely on modeled scenarios and have not yet been validated through production deployments. Conclusion: The modeled cost advantage appears to arise primarily from reducing duplicated and inconsistent governance processes rather than from eliminating security controls. Recommendations: Organizations operating regulated database workloads across two or more public clouds should prioritize federated identity management, policy-as-code security baselines, centralized telemetry, and automated audit-evidence generation before expanding provider-specific security tooling.

Energy-Aware Digital Twin Allocation with LLMs for Adaptive Resource Management in Cloud-Native Infrastructures

Managing containerized workloads in cloud-native infrastructures poses complex challenges due to the need to simultaneously balance performance, efficiency, and sustainability. This work proposes an adaptive resource allocation framework that leverages Digital Twins for real-time system monitoring and integrates Large Language Models to support context-aware decision-making under multi-objective constraints. The proposed approach dynamically optimizes latency, bandwidth utilization, and energy consumption, enabling intelligent workload orchestration across heterogeneous data center environments. A flexible utility function is introduced to allow system operators to adjust trade-offs between responsiveness and environmental impact. Experimental results demonstrate that the framework consistently outperforms traditional heuristic and learning-based baselines, achieving higher allocation accuracy, improved network utilization, and faster workload completion, while reducing overall energy consumption by more than 20% in sustainability-oriented scenarios. These findings highlight the potential of combining digital twins-driven observability with large language model-based reasoning to enable interpretable, adaptive, and energy-efficient resource management in next-generation cloud computing environments.