就Spring引导应用程序而言，bean只是一个Java对象，它是在应用程序启动时由Spring框架创建的。

对象的用途可以是任何东西——配置、服务、数据库连接工厂等等——Spring并不关心这些。

大多数bean依赖于其他bean才能工作，例如实体管理器可能需要一个数据库连接。Spring框架能够找出如何将bean自动连接在一起。从应用程序开发人员的角度来看，您只需声明所需的bean，它们就会“神奇地”出现在您的应用程序中，随时可以使用。

For a Java class to be usable as a Java bean, its setter- and getter-method names need to be as per the JavaBean guidelines (also called design patterns) for properties. If such a Java class is instantiable & manageable by the Spring IoC container, it is a Spring bean. To achieve this, the programmer wires the class as a bean definition of a suitable scope by using XML config files or annotations or a mix of both. The programmer can create new Spring beans out of existing Spring beans by wiring further by passing the latter to constructor-arguments of the former either as string-names as <idref> elements or by dependency injection (it can be recursive).

这个答案可以与我的这个SO答案一起阅读，以获得更多的背景信息。

Spring beans are just object instances that are managed by the Spring IOC container. Spring IOC container carry the Bag of Bean.Bean creation,maintain and deletion are the responsibilities of Spring Container. We can put the bean in to Spring by Wiring and Auto Wiring. Wiring mean we manually configure it into the XML file. Auto Wiring mean we put the annotations in the Java file then Spring automatically scan the root-context where java configuration file, make it and put into the bag of Spring.

For Spring, all objects are beans! The fundamental step in the Spring Framework is to define your objects as beans. Beans are nothing but object instances that would be created by the spring framework by looking at their class definitions. These definitions basically form the configuration metadata. The framework then creates a plan for which objects need to be instantiated, which dependencies need to be set and injected, the scope of the newly created instance, etc., based on this configuration metadata. The metadata can be supplied in a simple XML file, just like in the first chapter. Alternatively, one could provide the metadata as Annotation or Java Configuration.

书:Just Spring

在Spring中，那些构成应用程序主干并由Spring IoC管理的对象 容器被称为bean。bean只是一个实例化、组装等的对象 由Spring IoC容器管理;

Bean是一个POJO(普通旧Java对象)，由spring容器管理。

默认情况下，Spring容器只创建bean的一个实例。
此bean被缓存在内存中，因此对该bean的所有请求将返回对同一bean的共享引用。

@Bean注释返回一个对象，spring在应用程序上下文中将该对象注册为bean。
方法内部的逻辑负责创建实例。

什么时候使用@Bean注释?

当没有自动配置选项时。 例如，当我们想从第三方库连接组件时，因为源代码不可用，所以我们不能用@Component注释类。

实时场景可能是某人想要连接到Amazon S3桶。 因为源不可用，所以他必须创建一个@bean。

@Bean
public AmazonS3 awsS3Client() {
    BasicAWSCredentials awsCreds = new BasicAWSCredentials(awsKeyId, accessKey);
    return AmazonS3ClientBuilder.standard().withRegion(Regions.fromName(region))
            .withCredentials(new AWSStaticCredentialsProvider(awsCreds)).build();
}

源代码以上的代码-> https://www.devglan.com/spring-mvc/aws-s3-java

因为我上面提到了@Component Annotation。

@Component表示带注释的类是一个“组件”。在使用基于注释的配置和类路径扫描时，这些类被认为是自动检测的候选类。

组件注释将类注册为单个bean。