The main goal of the Amber Data Repository (ADR) is to analyse and share resilience measurement data, including field data on failures in real systems and experiment results. ADR integrates data from different sources in such way that enables comparison and cross-exploitation in a meaningful manner. In fact, this repository is a powerful tool to share data and intends to be a very important source of information for many research groups in Europe.
The ADR is part of the Work Package Coordination and Information Exchange Platform (WP1) of AMBER, an European Coordination Action started in January 2008. AMBER aims to coordinate the study of resilience measuring and benchmarking in computer systems and components, fostering European research in order to address the big challenges on resilience assessment posed by current and forthcoming computer systems and computer-based infrastructures. Further details about AMBER project is available at www.amber-project.eu.
Amber Data Repository has been developed under a business intelligence approach constituted of four main perspective: Data warehouse to store experimental results; On-Line Analytical Processing (OLAP) applicatons to analyze the results (statistical and ad-hoc analysis of multidimensional data); Data mining algorithms to identify (unknown) potentially interesting phenomena in the data; Information retrieval for raw data heavily base on text and XML.These perspectives may be used to query the ADR content through the option Explore Data.
AMBER Data Repository Team
Chalmers University of Technology submitted raw data on partitioning kernel using fault injection
University of Coimbra submited raw data on database dependability benchmarking
