Understanding Spring Bean Instantiation Process

Spring's IoC container works like a production line: it first prepares configuration metadata, then creates and manages bean instances. The whole bean creation can be divided into two major phases: the container startup phase and the bean instantiation phase.

Container Startup Phase

During this phase Spring loads configuration metadata from XML, properties files, annotations, or code, and represents each bean definition as a BeanDefinition . Different BeanDefinitionReader implementations (e.g., XmlBeanDefinitionReader , PropertiesBeanDefinitionReader , AnnotatedBeanDefinitionReader ) parse the source and register the resulting BeanDefinition objects in a BeanDefinitionRegistry .

Before any bean is instantiated, BeanFactoryPostProcessor can modify the registered definitions, for example replacing placeholder values:

<bean id="dataSource" class="org.apache.commons.dbcp.BasicDataSource" destroy-method="close">
  <property name="url" value="${jdbc.url}"/>
  <property name="username" value="${jdbc.username}"/>
  <property name="password" value="${jdbc.password}"/>
</bean>

Bean Instantiation Phase

When an application requests a bean, Spring creates the object using a strategy pattern: either simple reflection or a CGLIB proxy. The created instance is wrapped in a BeanWrapper , which simplifies reflective property access.

After creation, Spring injects properties. Primitive values are set directly from the definition; reference properties are resolved from already‑created beans, and if a dependency is missing Spring temporarily pauses the current bean, creates the required dependency, then resumes.

Spring also checks for Aware interfaces (e.g., ApplicationContextAware , ResourceLoaderAware ) and injects the corresponding container objects.

BeanPostProcessor Extensions

During the bean lifecycle, BeanPostProcessor runs both before and after custom initialization. It enables features such as AOP proxy creation and allows developers to modify or replace bean instances.

Custom initialization can be defined by implementing InitializingBean or specifying an init-method . Destruction logic can be provided via DisposableBean or a destroy-method (e.g., closing a datasource).

Summary of Execution Order

The overall order is: load metadata → register BeanDefinition → apply BeanFactoryPostProcessor → create bean instance → wrap with BeanWrapper → inject properties → invoke Aware callbacks → run BeanPostProcessor (pre‑ and post‑) → execute init methods → bean ready for use. When the container shuts down, destroy callbacks are invoked.