Building Efficient, Native Storage Systems for Semi-structured Data

IIS-0534530 Raju Rangaswami raju@cis.fiu.edu（PI）Vagelis Hristidis vagelis@cis.fiu.edu（Co-PI）佛罗里达国际大学计算与信息科学学院标题：为半结构化数据构建高效的本地存储系统摘要：该项目的目标是为存储和访问半结构化数据提供通用的、高效的本地支持。实现这一目标的第一步是通过开发将其结构紧密映射到常见存储设备（硬盘和基于MEMS的存储）的方法来最大限度地减少导航半结构化数据的访问开销。第二步是增强已开发的存储机制和接口，以允许有效实现面向数据的关键操作（锁定、访问权限实施、索引创建和有效更新）。第三步是设计并实现一个原生半结构化存储系统原型。 在本项目过程中开发的对半结构化数据的原生支持将对多个应用领域产生广泛影响，包括XML、传统目录文件结构、视频数据库、生物信息学数据分析以及广泛处理半结构化数据的其他物理科学领域。此外，方案执行机构还让来自少数族裔社区的佛罗里达国际大学学生参与该项目。这将有助于提高少数民族学生的毕业率，特别是博士学位。因此，他们在研究界的代表性。该项目的结果将通过项目网址http://www.cis.fiu.edu/SSS/传播。

IIS-0534530Raju Rangaswami raju@cis.fiu.edu (PI)Vagelis Hristidis vagelis@cis.fiu.edu (Co-PI)School of Computing and Information SciencesFlorida International UniversityTitle: Building Efficient, Native Storage Systems for Semi-structured DataAbstract:The objective of the proposed project is to provide generic, efficient, and native support for storing and accessing semi-structured data. The first step towards this goal is to minimize the access overhead for navigating semi-structured data by developing methodologies that tightly map its structure to those of common storage devices (hard disks and MEMS-based storage). The second step is to augment the developed storage mechanisms and interface to allow efficient implementation of key data-oriented operations (locking, access rights enforcement, index creation, and efficient updates). The third step is to design and implement a prototype of the native semi-structured storage system. Native support for semi-structured data developed during the course of this project will result in broad impact across several application domains including XML, legacy directory-file structures, video databases, Bioinformatics data analysis, and other areas of physical sciences that deal extensively with semi-structured data. Furthermore, the PIs are involving Florida International University students from minority communities in the project. This will help increase the graduation rate of minority students especially at the Ph.D. level, and consequently their representation in the research community. The results of the project will be disseminated via the project Web site: http://www.cis.fiu.edu/SSS/.