Jena-HBase: A Distributed, Scalable and
Efficient RDF Triple Store
Vaibhav Khadilkar
1
, Murat Kantarcioglu
1
, B havani Thuraisingham
1
, and
Paolo Castagna
2
1
The University of Texas at Dallas
2
Talis Systems Ltd.
Abstract. Lack of scalability is one of the most significant problems
faced by single machine RDF data stores. The advent of Cloud Comput-
ing has paved a way for a distributed ecosystem of RDF triple stores that
can potentially allow up to a planet scale storage along with distributed
query processing capabilities. Towards this end, we present Jena-HBase,
a HBase backed triple store that can be used with the Jena framework.
Jena-HBase provides end-users with a scalable storage and querying so-
lution that supports all features from the RDF specification.
1 Introduction
The simplest way to stor e RDF triples comprises a re lation/table of three columns,
one each for subjects, predicates and objects. However, this approach suffers from
lack of scalability and abridged query perfor mance, as the sing le table becomes
long and narrow when the number of RDF triples increases [1]. The approach is
not scalable since the table is usually located on a single machine. In addition,
query performance is diminished since a query requires several self-joins with the
same table. There have been several approaches, for e.g. [2], which modify the
single-table storage schema to improve query performance. Nevertheless, these
approaches suffer from the scalability proble m, which can be solved by moving
from a single-machine to a multi-machine configuration. The cloud computing
paradigm has made it possible to harness the processing power of multiple ma-
chines in parallel. Tools such as Hadoop and HBase provide advantages such as
fault tolerance and o ptimizations for r eal time queries. In this paper, we present
Jena-HBase
3
, a HBase backed triple store that can be used with the Je na frame-
work along with a preliminary expe rimental evaluation of our proto type.
Our work focuses on the creation of a distributed RDF storage framework,
thereby mitigating the scalability issue that exists with s ingle-machine systems.
The motivation to opt for Jena is its w ide spread acceptance, and its built-in sup-
port for manipulating RDF data as well as developing ontologies. Further, HBase
was selected for the storage layer for two reasons: (i) HBase is a column-oriented
store and in general, a column-oriented store performs better than row-oriented
stores [1]. (ii) Hadoop comprises Hadoop Distributed File System (HDFS), a dis-
tributed file system that stores data, and MapReduce, a framework for processing
data stored in HDFS. HBase uses HDFS for data storage but does not require
MapReduce for accessing data. Thus, Jena-HBase does not require the imple-
mentation of a MapReduce-based query engine for executing queries on RDF
3
https://github.com/castagna/hbase-rdf, https://github.com/vaibhavkhadilkar/hbase-rdf
下载后可阅读完整内容，剩余3页未读，立即下载
白乔
- 粉丝: 3423
- 资源: 14
我的内容管理
展开
最新资源
- WebLogic集群配置与管理实战指南
- AIX5.3上安装Weblogic 9.2详细步骤
- 面向对象编程模拟试题详解与解析
- Flex+FMS2.0中文教程：开发流媒体应用的实践指南
- PID调节深入解析：从入门到精通
- 数字水印技术：保护版权的新防线
- 8位数码管显示24小时制数字电子钟程序设计
- Mhdd免费版详细使用教程：硬盘检测与坏道屏蔽
- 操作系统期末复习指南：进程、线程与系统调用详解
- Cognos8性能优化指南：软件参数与报表设计调优
- Cognos8开发入门：从Transformer到ReportStudio
- Cisco 6509交换机配置全面指南
- C#入门：XML基础教程与实例解析
- Matlab振动分析详解：从单自由度到6自由度模型
- Eclipse JDT中的ASTParser详解与核心类介绍
- Java程序员必备资源网站大全
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
