CAM2S: An Integrated Indexing Structure for Spatial Objects Generating Data Streams

Data warehouse systems store and process huge amounts of data. The main application of data warehouses is to support decision making process and to provide reports of various kinds. Those tasks require calculating aggregates, averages and other values. The calculations must be performed fast because it is very common that user wants to work interactively with a data warehouse system. There are two ways to increase the data warehouse performance. The first is to increase computing power by improving hardware infrastructure, and the second, to apply advanced, dedicated solutions like materialization, indexing etc. In this paper we present an indexing structure designed for spatial telemetric objects and suitable for processing streams of data. The presented solution is a result of integration of a few mechanisms. For spatial indexing we use CHR-Tree index which idea is based on the R-Tree index. To process data streams we use Materialized Aggregate List, a solution which allows processing data streams of any length and supports optional stream materialization. In the paper we present how those solutions are integrated into one indexing and caching structure. We also include efficiency test results showing the solution efficiency and scalability.

I. Introduction

In recent years data warehouse systems become more and more popular. Users find the ability of processing large amounts of data in a short time very useful and convenient. This trend is supported by products of large software companies who enrich their business offer with ready-to-use data warehouse solutions integrated with broadly known database systems. A simple data warehouse can be created by a user who, by clicking, models data warehouse structure, defines dimensions, attributes and hierarchies and finally defines what reports should be available for a final user. Systems build from a scratch and dedicated for a single company are no longer the only way to create a data warehouse.