不一样的开源DBMS iMeMex

关键点：引入了“数据空间”这个新概念，突破了传统多数据模型数据源的处理方法，一切都“关系”的管理理念有待发展，依靠集成与转换的传统处理方法，灵活性与可控性不好，所以现有的基础理论与技术研究一定要发展，这是需求驱动的结果，IBM  DB2 9已经有双核DBMS的实践，内核的多数据模型支持要比外围的转换与集成有更好的性能，非功能性要求表现更好，这方面很值得探索，可以借助开源的先驱力量进行再探索。比如 iMeMex
http://imemex.ethz.ch/
Dr. Jens-Peter Dittrich (project lead, Senior Research Associate)

Abstract

I am a Senior Research Associate (in German: `Oberassistent´) in the Databases and Information Systems Group (DBIS) of Prof. Donald Kossmann, ETH Zurich. My primary research interests include Personal DataSpace Management (iMeMex), efficient algorithms for large datasets (joins, indexing), as well as algorithms for the efficient analysis of large data repositories (OLAP, personal information spaces).

My research has a strong focus on system building: during my Ph.D. I coauthored the open source project XXL - a library of advanced query processing algorithms. After that, I worked for SAP's OLAP engine (SAP BW) and SAP's enterprise search engine (TREX) on a joint project on high performance analytics (BI Accelerator). Since late 2004, I have been working on the Personal DataSpace Management System iMeMex (supported by the Swiss National Science Foundation (SNF) with a 3 year fund).