From Local to Remote: VisIVO Visual Analytics in the Era of the Square Kilometre Array

Giuseppe Tudisco, Fabio Vitello, Eva Sciacca, Ugo Becciani · 2025

Read From Local to Remote: VisIVO Visual Analytics in the Era of the Square Kilometre Array by Giuseppe Tudisco, Fabio Vitello, Eva Sciacca, Ugo Becciani (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-24113.txt.

Introduction

A giant radio telescope now being built will pour out hundreds of petabytes of sky data each year. Giuseppe Tudisco, Fabio Vitello, Eva Sciacca, and Ugo Becciani describe a viewing tool for that flood in a 2025 paper. Their tool, VisIVO, helps astronomers see patterns in huge data cubes.

The tool is moving from a desktop program to a client-server system. Heavy work will run in parallel on computing clusters, while the user steers from a light front end. That shift lets teams open and probe files too large for one machine. The authors frame the move as a must for the new telescope era, where data size outgrows local disks.

Catalog record

CREATORGiuseppe Tudisco, Fabio Vitello, Eva Sciacca, Ugo Becciani
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE12 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-24113

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

From Local to Remote: VisIVO Visual Analytics in the Era of the Square Kilometre Array Giuseppe Tudisco, Fabio Vitello, Eva Sciacca, Ugo Becciani Abstract The field of astrophysics is continuously advancing, with an ever-growing influx of data requiring robust and efficient analysis tools. As the Square Kilometre Array (SKA) radio telescopes come fully operational, we anticipate the generation of hundreds of petabytes of data annually, characterized by unprecedented resolution and detail. In this context, scientific visualization becomes a critical component, enabling researchers to interpret complex datasets and extract meaningful insights. The immense volume of data demands not only suitable tools but also substantial infrastructure and computational capacity to analyze it effectively. In this work, we will discuss how we are addressing these challenges with the development of our interactive visualization tool named VisIVO Visual Analytics. The tool is transitioning from a local visualizer to a remote visualizer, utilizing a client-server architecture. This evolution will allow the software to run parallel visualization pipelines on high-performance computing (HPC) clusters, thereby enhancing its capacity to handle extensive datasets efficiently. 1 Introduction Scientific visualization plays a pivotal role in translating complex datasets into meaningful insights, helping researchers comprehend intricate astrophysical phenomena. The main goal of these visualization techniques is to provide tangible outputs, such as images, 3D models, and animations, which allow researchers to explore subtle properties and uncover relationships hidden within vast and complex datasets. Beyond mere representation, advanced visualization tools enable in-depth investigations of aspects that would otherwise be challenging to analyze. As astrophysical datasets are growing in both scale and complexity – particularly with the forthcoming Square Kilometre Array (SKA) expected to generate a staggering 700 petabytes of processed data annually – the development of more sophisticated visualization tools becomes essential for managing and analyzing such massive datasets. The sheer volume and complexity of data produced by state-of-the-art astronomical instruments necessitate the creation of advanced algorithms, computational techniques, and storage solutions capable of handling these ever-increasing demands. 2 VisIVO Visual Analytics overview Since 2005, the INAF Astrophysical Observatory of Catania has been developing and maintaining the Visualization Interface for the Virtual Observatory (VisIVO), a set of tools designed to facilitate 3D and multidimensional data analysis and visualization. VisIVO enables researchers to uncover previously unknown relationships within complex, multivariate astrophysical datasets, supporting advanced knowledge discovery. VisIVO Visual Analytics [ 4 ] , part of the VisIVO family, is a powerful tool designed to provide an interactive environment for visualizing and analyzing FITS images and cubes. Through its 3D datacube representation, the Visual Analytics tool allows users to systematically explore complex datasets, offering a comprehensive view of data distribution (see Figure 1 ). This capability aids in detecting structures and gradients within the three-dimensional space, making it a useful tool for investigating intricate astrophysical phenomena. Figure 1: Visualization of a datacube in VisIVO Visual Analytics. The volume rendering is shown on the left, the selected slice is shown on the right. VisIVO also offers a variety of analytical tools for detailed exploration of datacubes, including the generation of Position-Velocity diagrams, spectral axis analysis, and calculation of moment maps. These features enable scientists to extract critical information from datacubes, such as velocity profiles, intensity distributions, and velocity dispersions. Beyond handling local files, VisIVO Visual Analytics can search and retrieve the extensive resources of the ViaLactea Knowledge Base [ 2 ] through dedicated Search and Cutout services, allowing users to access over 40,000 datasets organized into nearly 100 collections, with a Table Access Protocol service (TAP) providing access to approximately 15 source catalogs. Users can easily interact with the TAP service through the tool’s intuitive interface, which allows them to define a rectangular image area to specify the region of interest and execute queries for compact sources and filamentary structure catalogs. Additionally, researchers can perform Spectral Energy Distribution (SED) analysis and fitting operations on compact sources overlaid on images. This broad functionality enhances the tool’s versatility, enabling researchers to incorporate external datasets and resources, ultimately expanding the potential for scientific inquiry and analysis. 2.1 Remote Visualization in VisIVO To meet the unique challenges posed by the Square Kilometre Array, efforts are underway to design and implement a remote visualization model within the VisIVO Visual Analytics for visualizing data stored on a separate computational system from the one used for image rendering. Remote visualization for large astrophysical datasets provides several key benefits. This approach not only improves efficiency but also broadens accessibility. By separating data processing from visualization, it optimizes the use of computational resources and overcomes the limitations of common personal hardware. This approach is being integrated into VisIVO to address the difficulties of managing, accessing, and analyzing massive astrophysical datasets, such as the ones expected from the SKA Observatory. The goal is to develop innovative solutions that enhance both data management and analytical capabilities, ensuring efficient handling of these large-scale datasets. Over the past few years, we have explored and implemented various remote visualization solutions to address this aspect. ViaLactea Web Our initial approach was ViaLactea Web [ 3 ] , a web application that offered basic VisIVO Visual Analytics functionality, such as 2D images and 3D cubes visualization (stored on the server), isocontour rendering, and the ability to search and download cutouts from the ViaLactea Knowledge Base. The frontend was built with Node.js, while the backend was developed in Python, with some visualization pipeline code written in C++. These components communicated with each other via web sockets. However, this solution required extensive rewriting of the original VisIVO Visual Analytics codebase to be adapted in a web environment. Figure 2: ViaLactea Web running on a server and visualizing a datacube. Containerization of VisIVO in a VNC environment To preserve the full functionality of the existing application while enabling remote data visualization, we developed a VisIVO Visual Analytics container based on an Ubuntu image with a Desktop Environment installed. Thanks to the NVIDIA container toolkit 1 1 1 https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/ , TurboVNC 2 2 2 https://turbovnc.org , and VirtualGL 3 3 3 https://virtualgl.org , the container can be deployed on a remote server with GPU support, allowing for hardware-accelerated and high-performance visualization. With appropriate volume mounting, this solution enables users to access and analyze remote datasets through a browser, offering the complete feature set of VisIVO Visual Analytics. Unlike ViaLactea Web, this container-based approach allows users to interact with remote data seamlessly while maintaining the tool’s full capabilities. Science Platform The next advancement was deploying VisIVO Visual Analytics through Science Platforms, specifically for the CANFAR Science Platform infrastructure and architecture as part of the work developed for the SKA Regional Centres (SRC). The CANFAR Science Platform runs on top of a Kubernetes cluster and allows its users to run application containers (desktop applications, web applications or Jupyter Notebook) in interactive user sessions. We published a VisIVO container image adapted for running on the CANFAR Science Platform (see Figure 3 ). The desktop sessions are made accessible through noVNC. Users can customize their sessions by specifying the amount of memory and CPU resources needed for the application. All sessions share a common storage system, ensuring persistence of data and settings across sessions, which enhances both the flexibility and scalability of the tool within the Science Platform. This

Read the complete text