分布式查询语言测试系统设计与实现

发布时间：2018-02-26 01:20

本文关键词： 分布式查询语言蝴蝶模型面向对象测试测试自动化　出处：《北京交通大学》2012年硕士论文　论文类型：学位论文

【摘要】：对搜索引擎用户日志的分析有利于提高搜索服务质量、发现潜在用户、增强竞争力。为完成海量日志数据分析,需要为日志分析平台构建分布式计算框架DISQL (Distribute Query Language)。本文基于分布式计算框架实际的测试需求,根据Hadoop Map-reduce分布式计算模型的特点,在对分布式查询语言分析的基础上,设计了DISQL测试系统。该系统包含四个独立测试功能的模块,应用多种自动化测试技术实现了自动化测试,完成了对分布式查询语言的分布式测试。论文工作主要包括：深入挖掘分布式查询语言的测试功能点,详细分析DISQL测试系统的功能划分及测试的范围。根据Hadoop分布式计算系统以及分布式查询语言使用面向对象编程等特点,考虑到测试需求变更多、设计开发周期短等情况,分布式查询语言测试系统选用高效便捷的软件测试模型一蝴蝶模型,采用面向对象的软件测试技术,实现Linux操作系统中的自动化测试方案。基于测试用例的应用环境,结合分布式查询语言的系统结构,应用自顶向下的设计方法,设计了对测试分布式查询语言的翻译器与算子进行集成测试的API算子测试模块；设计了对分布式查询语言用户接口正确性进行验证的SDK测试模块；设计了可衡量分布式计算框架的处理性能的性能测试模块；设计了可对用户自定义算子进行模块测试的C回调测试模块。针对各模块中不同测试场景的测试用例,结合测试用例的逻辑及功能特点,应用PHP面向对象编程技术,综合使用测试脚本模块化框架、测试库架构框架、数据驱动测试框架等结构等自动化框架,编码实现各模块的自动化测试框架并将测试用例逐步集成到DISQL测试系统。应用DISQL测试系统,在Hadoop集群上对分布式查询语言进行迭代的回归测试,经测试,DISQL测试系统可实现对分布式语言系统集成测试和系统测试,为DISQL的顺利发布提供重要支持,同时为DISQL'性能升级提供重要参考指标。
[Abstract]:The analysis of search engine user log is helpful to improve the quality of search service, find potential users and enhance competitiveness. It is necessary to build a distributed computing framework, DISQL distributed Query language, for log analysis platform. Based on the actual test requirements of the distributed computing framework and the characteristics of the Hadoop Map-reduce distributed computing model, this paper analyzes the distributed query language. The DISQL test system is designed. The system includes four modules with independent test function. The automatic test is realized by using a variety of automatic test techniques. The distributed testing of distributed query language is accomplished. The main work of this paper is as follows:. The test function points of distributed query language are deeply mined, and the function partition and test range of DISQL test system are analyzed in detail. According to the characteristics of Hadoop distributed computing system and distributed query language using object-oriented programming, etc. Considering that the testing requirements are changing and the design and development cycle is short, the distributed query language testing system adopts the efficient and convenient software testing model-Butterfly model, and the object-oriented software testing technology. The automatic test scheme in Linux operating system is realized. Based on the application environment of test cases, combining the system structure of distributed query language and the top-down design method, the API operator test module is designed to integrate the translator and operator of the test distributed query language. The SDK test module is designed to verify the correctness of the distributed query language user interface, and the performance test module to measure the processing performance of the distributed computing framework is designed. A C callback test module is designed to test the user defined operator. According to the test cases of different test scenarios in each module, combined with the logic and function characteristics of test cases, PHP object-oriented programming technology is applied, the modular framework of test scripts and the framework of test library architecture are used synthetically. The automation framework, such as data-driven test framework and so on, is coded to realize the automated test framework of each module and the test cases are integrated into the DISQL test system step by step. Using the DISQL test system, the distributed query language is tested iteratively on the Hadoop cluster. By testing the DISQL test system, the distributed language system integration test and system test can be realized, which provides important support for the smooth release of DISQL. At the same time for DISQL 'performance upgrade to provide an important reference index.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.3

【参考文献】