Collaborative Research: Query Optimization Engineering

这是俄勒冈科学技术研究生院（OGI）和波特兰州立大学（PSU）之间的一个机构间合作项目。查询优化器是现代数据库系统实现其性能优势的主要手段之一。给定一个数据操作或检索请求，优化器将从多种可选策略中选择一个最优计划来评估该请求。基本关系模型的优化被认为是一个已解决的问题。然而，新的应用领域，如知识发现，在线分析处理和复杂的多媒体对象，已经点燃了新的研究优化。目前的优化器往往被证明不足以满足这些新应用领域的需要。研究人员已经开发了优化技术和几个原型优化器，以支持新的应用。这项研究解决了与在新应用中使用新优化器相关的工程问题：使用哪些操作器；采用哪种搜索策略；然后转化为利用。这些调查使用了PSU和OGI开发的Columbia查询优化器框架和可视化环境。该项目与生产数据库管理系统及其运行硬件的公司密切互动。研究结果将有助于数据库优化器实现者在设计系统时做出更好的选择，从而为商业数据库系统提供更好的性能和更具表现力的查询语言。

This is an interinstitutional collaborative project between Oregon Graduate Institute of Science and Technology (OGI) and Portland State University (PSU). Query optimizers are one of the main means by which modern database systems achieve their performance advantages. Given a request for data manipulation or retrieval, an optimizer will choose an optimal plan for evaluating the request from among the manifold alternative strategies. Optimization for the basic relational model is considered a solved problem by many. However, new application areas such as knowledge discovery, on-line analytical processing, and complex multimedia objects, have kindled renewed research in optimization. Current optimizers have often proved inadequate to the needs of these new application areas. Researchers have developed optimization techniques, and several prototype optimizers, to support new applications. This research addresses engineering questions related to the use of new optimizers for new applications: which operators to use; which search strategies to employ; and which transforms to utilize. These investigations use the Columbia query optimizer framework and visualization environment developed at PSU and OGI. The project interacts closely with companies which produce database management systems and the hardware on which they run. The research results will help database optimizer implementors make better choices in the design of their systems, leading to better performance and more expressive query languages for commercial database systems.