ShareX3D, a scientific collaborative 3D viewer over HTTP

Abstract

International audienceIn scientific visualization data are becoming more and more complex and implies cooperative effort for their post-processing analysis as well as high performance processing resources. Moreover experts are frequently geographically distributed and existing post-processing tools do not provide high performance capability with collaborative features. To solve this limitation, a prospective action has been recently initiated in the frame of the ANR SCOS project, with the development of a web based post-processing framework called V3D. SCOS/V3D is based on two innovating concepts: the sharing of the processing results via the transfer of a X3D data file corresponding to the final virtual reality scene and communications based on HTTP to by-pass standard proxies and firewalls limitations. To test this approach, a proof of concept prototype, called shareX3D, has been developed, especially focusing on the events notification, with an implementation of the long polling technique. The architecture of ShareX3D and first tests are presented and discusse

Streaming compressed 3D data on the web using JavaScript and WebGL

“…Many works on web-based remote visualization of 3D data have been conducted for remote collaborative scientific visualization for which an overview has recently been conducted by Mouton et al [2011]. ShareX3D [Jourdain et al 2008] provides a web-based remote visualization of 3D data; the web rendering is based on Java. The COVISE (COllaborative VIsualization and Simulation Environment) platform, in its last version [Niebling and Kopecki 2010], offers a web client implemented using JavaScript and WebGL.…”

“…The COVISE (COllaborative VIsualization and Simulation Environment) platform, in its last version [Niebling and Kopecki 2010], offers a web client implemented using JavaScript and WebGL. However these systems, like most of existing ones, make use of XML-based ASCII format such as VRML or X3D [Jourdain et al 2008] or JavaScript vertex arrays [Niebling and Kopecki 2010] to exchange the 3D data, which involves a significant latency due to the large file size. To resolve this latency issue, some compression methods have been proposed.…”

A hybrid visualization system for molecular models

“…Data management remains a critical part of the system, especially when the datasets are very large and contain sensitive information. To leverage the flexibility of cloud-based solutions, we used the Midas Platform [Jourdain et al 2008], an open-source toolkit that enables the rapid creation of tailored, web-enabled data storage. The Midas Platform addresses the growing challenge of large data by providing a flexible, intelligent data storage system.…”

“…However, as the data gets larger, it becomes challenging to obtain real-time visualization on a personal computer. In the past three years, several efforts have been trying to solve this problem using varying technologies such as ParaViewWeb [Jourdain et al 2010] and [Jomier et al 2011], ShareX3D [Jourdain et al 2008] or Collaviz Framework [Dupont et al 2010]. All these projects have the common goal to bring visualization to the user via a thinclient, therefore pushing the computational resources to the backend server.…”

Getting ready for petaflop capacities and beyond: a utility perspective

“…We are currently putting together a 6 by 3 meter, 23 Mpixels Power Wall, together with parallel visualization tools and clusters, in order to boost our visualization capacities of the most advanced simulations. However, the sheer volume of data produced by petaflop calculations, storage and network limitations, and multisite teams make it necessary to further advance R&D in a number of areas [12]. Under the CARRIOCAS project and the SCOS project we have started work on remote and collaborative multiuser visualization, parallel visualization, and parallel and distributed file systems.…”