DataLab as a Service: Distributed Computing Framework for Multi-Interactive Analysis Environments

作     者：Hoz, Aida Palacio Heredia Canales, Andres Cimadevilla Alvarez, Ezequiel Obregon Ruiz, Marta Lopez Garcia, Alvaro 

作者机构：Inst Fis Cantabria IFCA CSIC UC Santander 39005 Spain 

出 版 物：《IEEE ACCESS》 (IEEE Access)

年 卷 期：2025年第13卷

页      面：22566-22577页

核心收录：

基　　金：European Open Science Cloud (EOSC) SIESTA Project "Secure Interactive Environments for Sensitive daTa Analytics" - European Union (Horizon Europe) 

主　　题：Cloud computing Microservice architectures Ecosystems Computer architecture Servers Scalability Libraries Complexity theory Codes Sparks Distributed computing Kubernetes interactive environments cloud computing Jupyter notebooks 

摘      要：During the current data era, data analysis across multiple disciplines has become a critical task for researchers to obtain meaningful insights and solve complex problems that are immeasurable using traditional technologies. Big Data has led to the development of state-of-the-art technologies that have revolutionized the process of experimentation. These innovations span from automating the setup of the infrastructure required for data analysis to providing user-friendly interfaces that simplify coding and result visualization. However, managing and scaling these resources for large-scale data processing remains a challenge. In this work, we introduce a novel framework called Datalab as a Service which integrates cutting-edge and open-source technologies to offer an online platform designed for both resource providers and researchers. The platform enables users to easily and automatically deploy interactive environments tailored for data analysis, thereby streamlining the process of managing computational resources. Through DLaaS, users gain access to cloud-based infrastructures and distributed computing resources, which are essential for performing compute-intensive tasks on massive datasets. The framework ensures scalability, resource management and optimization, and high availability, all within an accessible and user-friendly platform. Furthermore, this paper presents several use cases where researchers have successfully utilized DLaaS resources, demonstrating its practical applications in real-world scenarios.