Mastering Database Read/Write Splitting with Spring AOP: A Practical Demo

Introduction

Read/write splitting in a distributed environment is a key solution to database read/write performance bottlenecks, greatly improving read speed and concurrency.

To achieve it, a master‑slave configuration is required. The simplest setup uses one Master and one Slave; reads go to the Slave, writes to the Master, reducing load on a single server.

Two Implementation Methods

Method 1: Define two data sources, MasterDataSource for writes and SlaveDataSource for reads. The application explicitly selects the appropriate source.

Method 2: Dynamic data‑source switching using AOP. At runtime the program weaves the data source based on the invoked method, leveraging annotations, Spring AOP and reflection.

AOP‑Based Read/Write Splitting Demo

Project code is available at the open‑source repository:

http://git.oschina.net/xuliugen/aop-choose-db-demo

Key components include:

ChooseDataSource – selects the appropriate data source.

DataSourceAspect – intercepts methods to determine read or write operations.

DataSourceHandler – handles the actual data‑source routing.

Configuration files (Spring XML) define readDataSource, writeDataSource, and a dynamic dataSource bean that references ChooseDataSource. The XML snippet is shown below:

<beans xmlns="http://www.springframework.org/schema/beans"
       xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
       xmlns:aop="http://www.springframework.org/schema/aop"
       xsi:schemaLocation="
           http://www.springframework.org/schema/beans
           http://www.springframework.org/schema/beans/spring-beans-4.0.xsd
           http://www.springframework.org/schema/aop
           http://www.springframework.org/schema/aop/spring-aop.xsd">

    <bean id="statFilter" class="com.alibaba.druid.filter.stat.StatFilter" lazy-init="true">
        <property name="logSlowSql" value="true"/>
        <property name="mergeSql" value="true"/>
    </bean>

    <!-- read data source -->
    <bean id="readDataSource" class="com.alibaba.druid.pool.DruidDataSource"
          destroy-method="close" init-method="init" lazy-init="true">
        <property name="driverClassName" value="${driver}"/>
        <property name="url" value="${url1}"/>
        <property name="username" value="root"/>
        <property name="password" value="${password}"/>
    </bean>

    <!-- write data source -->
    <bean id="writeDataSource" class="com.alibaba.druid.pool.DruidDataSource"
          destroy-method="close" init-method="init" lazy-init="true">
        <property name="driverClassName" value="${driver}"/>
        <property name="url" value="${url}"/>
        <property name="username" value="root"/>
        <property name="password" value="${password}"/>
    </bean>

    <!-- dynamic routing data source -->
    <bean id="dataSource" class="com.xuliugen.choosedb.demo.aspect.ChooseDataSource" lazy-init="true">
        <property name="targetDataSources">
            <map key-type="java.lang.String" value-type="javax.sql.DataSource">
                <entry key="write" value-ref="writeDataSource"/>
                <entry key="read" value-ref="readDataSource"/>
            </map>
        </property>
        <property name="defaultTargetDataSource" ref="writeDataSource"/>
        <property name="methodType">
            <map key-type="java.lang.String">
                <entry key="read" value=",get,select,count,list,query"/>
                <entry key="write" value=",add,create,update,delete,remove,"/>
            </map>
        </property>
    </bean>
</beans>

The demo uses the SSM framework (Spring, Spring MVC, MyBatis). The UserController simulates read/write operations by invoking IUserService. Diagrams illustrate the project structure and the flow of AOP‑based data‑source selection.