T2: A Customizable Parallel Database For Multi-dimensional Data

In this paper, we present T2, a customizable parallel databasethat integrates storage, retrieval and processing of multi-dimensionaldatasets. T2 provides support for common operations including indexgeneration, data retrieval, memory management, scheduling of processing acrossa parallel machine and user interaction. It achieves its primary advantagefrom the ability to seamlessly integrate data retrieval and processing for awide variety of applications and from the ability to maintain and jointlyprocess multiple datasets with different underlying grids. Most other systemsfor multi-dimensional data have focused on uniformly distributed datasets, suchas images, maps, and dense multi-dimensional arrays. Many real datasets,however, are non-uniform or unstructured. For example, satellite data consistsof a two dimensional strip that is embedded in a three dimensional space; watercontamination studies use unstructured meshes to selectively simulate regionsand so on. T2 can handle both uniform and non-uniform datasets.