首页> 外文学位 >A graph based cache system for efficient querying in distributed triplestores.
【24h】

A graph based cache system for efficient querying in distributed triplestores.

机译:基于图的缓存系统,可在分布式三元存储中进行有效查询。

获取原文
获取原文并翻译 | 示例

摘要

Predefining structures for data preclude the ability to describe certain details effectively, as data is dynamically structured by nature and can be highly diverse and multifaceted. Conventional datastores, such as relational databases, do not conveniently accommodate dynamically varying structures, as frequently modifying database schemas is not feasible. Although, XML databases have been proposed as suitable for such dynamic structured data, these databases suffer from update anomalies. RDF triplestores offer a flexible solution for handling such data, where any property about an entity can be described by a triple consisting of a subject, a predicate, and an object. Data is inherently distributed due to origination points, ownership, and many other reasons.;In this dissertation, we develop a distributed triplestore while investigating different approaches for improving the efficiency of query processing. We implement several index structures at a Mediator where each index structure helps to minimize unproductive communications. We show how the addition of each index structure reduces the query response time. We make use of graph-based caches at the Mediator and at individual triplestores to store triples that correspond to the most frequent set of queries. The Mediator cache enables the Mediator to respond to certain queries, while the triplestore caches avoid the need to perform joins between partial results from different triplestores. Sub-graph isomorphism is used to determine whether a cache has sufficient triples to answer a given query. We show how these caches significantly improve querying efficiency.
机译:数据的预定义结构排除了有效描述某些细节的能力,因为数据本质上是动态构造的,并且可以高度多样化和多方面。诸如关系数据库之类的常规数据存储无法方便地容纳动态变化的结构,因为频繁修改数据库模式是不可行的。尽管已经提出XML数据库适合于这种动态结构化数据,但是这些数据库存在更新异常的问题。 RDF三重存储为处理此类数据提供了灵活的解决方案,其中有关实体的任何属性都可以通过由主语,谓语和对象组成的三元形容来描述。由于起源,所有权和许多其他原因,数据固有地是分布式的;在本文中,我们研究了分布式三元组,同时研究了提高查询处理效率的各种方法。我们在中介器上实现了几个索引结构,每个索引结构都有助于最大程度地减少非生产性的沟通。我们展示了添加每个索引结构如何减少查询响应时间。我们在介体和各个三元存储中利用基于图的缓存来存储与最频繁查询集相对应的三元。 Mediator缓存使Mediator能够响应某些查询,而三重存储缓存则避免了在不同三重存储的部分结果之间执行联接的需要。子图同构用于确定高速缓存是否具有足够的三元组来回答给定查询。我们展示了这些缓存如何显着提高查询效率。

著录项

  • 作者

    Devadithya, Tharaka.;

  • 作者单位

    Indiana University.;

  • 授予单位 Indiana University.;
  • 学科 Computer Science.
  • 学位 Ph.D.
  • 年度 2008
  • 页码 186 p.
  • 总页数 186
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

  • 入库时间 2022-08-17 11:38:33

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号