REAL-WORLD BENCHMARK: Caché vs. Oracle in a Data Warehousing Application

Abstract

A benchmark of a real-world application, which loads data into a data warehouse for subsequent analysis, was performed. To conduct the benchmark, one module of the Oracle-based application was replicated in Caché ObjectScript. Only about 40 person-hours of work was required to duplicate the functionality of the original module in Caché.

The time required to complete certain tasks was measured running the Caché-based module, and was compared to existing data. The Caché application was five times faster, taking only 286 minutes to complete the data loading and filtering tasks that took 1441 minutes when using Oracle.

Introduction

Meralco Corporation, a large power supply company in the Philippines, currently uses Oracle for their corporate data warehouse. Every month a large amount of text-based data is loaded into the data warehouse, filtered, and made available for analysis. In an effort to improve their data warehousing operations, Meralco approached Digital Dimensions, Inc., a distributor of the high-performance Caché DBMS from InterSystems Corporation. Meralco agreed to let Digital Dimensions and InterSystems replicate the Field Order module of their Data Warehouse ETL (Extract, Transform, Load) application in a Caché-based system, in order to test Caché's performance compared to Oracle.