A Data Model for Distributed Multiresolution Multisource Scientific Data

Abstract

Summary.Modern dataset sizes present major obstacles to understanding and interpreting the significant underlying phenomena represented in the data. There is a critical need to support scientists in the process of interactive exploration of these very large data sets. Using multiple resolutions of the data set (multiresolution), the scientist can identify potentially interesting regions with a coarse overview, followed by narrower views at higher resolutions.Scientific data sets are often multisourcecoming from different sources. Although it may be infeasible to physically combine multiple datasets into a single comprehensive dataset, the scientist would often like to treat them as a single logical entity. This paper describes formal conceptual models of multiresolution and distributed multisource scientific data along with an implementation of our multisource model. Our goal is to allow a scientist to describe a dataset that combines several multisource multiresolution datasets into a single conceptual entity and to provide efficient and transparent access to the data based on functionality defined by the model.

A multiresolution data model for improving simulation I/O performance

“…We have developed a multiresolution scientific data model [1] that incorporates spatial and temporal semantics with localized error and we have implemented a database system based on that model [17], [18]. The model's semantics are common to many scientific applications and therefore are valuable to a variety of disciplines.…”

“…Second, the resolution limitation: both human and display resolution limits restrict the amount of information that can be displayed and processed at any moment. Ideally, we would The traditional multiresolution data representation [1], applied to visualizations, helps the scientist deal with very large datasets by offering a coarse resolution to see an overview of the data, and a finer resolution to see a more detailed animation at the cost of losing interactivity. We extend this model to the simulation environment in the following manner: we use high resolution during the simulation to reduce the numerical errors, but store the results on disk at lower resolutions.…”

Uncertainty Classification and Visualization of Molecular Interfaces

“…Uncertainty-driven multidimensional classification of scalar volume data has also been explored [26]. Uncertainty pertaining to isosurfaces with computed error has also been visualized [27]. Point-based approaches allow for expression of spatial uncertainty in both polygonal and volumetric data [28].…”

FIRST: a flexible and interactive resampling tool for CFD simulation data

“…To describe the algorithm, we introduce the notion of influence, similar to the notion introduced by Rhodes et al [12,13]. A triangle influences a resampler cell if the triangle contributes to the summary (or resampled) data of a structured grid cell.…”