基于AspectJ非侵入式采集Android应用性能指标研究  

Research on non⁃invasive collection of Android application performance indicatorsbased on AspectJ

作　　者：承轶青 贺敬伟 章路进 CHENG Yiqing;HE Jingwei;ZHANG Lujin(Jiangsu Electric Power Information Technology Co.,Ltd.,Nanjing 210024,China)

机构地区：[1]江苏电力信息技术有限公司,江苏南京210024

出　　处：《电子设计工程》2020年第21期164-167,172,共5页Electronic Design Engineering

基　　金：江苏省电力公司电力科学研究院科研项目(JGSQY00FCJS1800496)。

摘　　要：传统的应用性能指标研究方法需要使用自定义注解或者调用SDK侵入切入点代码,采集指标扩展依赖业务应用代码。为了解决上述问题,基于AspectJ研究了一种新的非侵入式采集Android应用性能指标的方法。研究性能指标采集方法开发插件,通过Gradle编译插件在编译过程中修改编译后的class文件内容,在切入点增加采集插件实现无埋点植入。通过采集插件,对An⁃droid应用性能指标数据进行预处理,进而上报存储数据。为了验证方法的有效性,设定对比实验,结果表明,基于AspectJ非侵入式采集Android应用性能指标研究方法可扩展性更强,业务代码无需改动,实用价值更高。Traditional research methods of application performance index need to use custom annotation or call SDK to invade the pointcut code,and collection index extension depends on business application code.In order to solve the above problems,a new method of non⁃invasive collection of Android application performance indicators is studied based on AspectJ.This paper studies the performance index collection method to develop plug⁃in,modifies the compiled class file content during the compilation process through the gradle compilation plug⁃in,and adds the collection plug⁃in at the entry point to realize the embedding without burying points.Through the collection plug⁃in,preprocess the Android application performance index data,and then report the stored data.In order to verify the effectiveness of the method,set up a comparative experiment,the results show that the research method based on AspectJ non⁃invasive collection Android application performance indicators has stronger scalability,no need to change the business code,and higher practical value.

关 键 词：ASPECTJ 非侵入式采集 Android应用性能 应用性能指标 

分 类 号：TN98[电子电信—信息与通信工程] TM615[电气工程—电力系统及自动化]