How to control OSGi dependency injection in AEM?

I'm trying to understand how dependency injection works in OSGi (specifically Apache Felix as used in AEM). I have a servlet with an `@Reference` annotation on a field that references an interface -- in my case, it represents a secure document signing provider. I have an implementation class that implements the interface, and it's automatically injected into the servlet.

In the servlet:

    @Reference(cardinality = ReferenceCardinality.MANDATORY_UNARY)

    private DocumentSigningProvider signingProvider;

    ...

    URL redirectUrl = signingProvider.Sign(...);

and my implementation class:

    @Component(metatype=true)

    @Service

    public class DocumentSigningDocuSignImpl implements DocumentSigningProvider {

    ...

    @Override

    public URL Sign(...) {...}

As you can see, in my servlet I code against the interface, not the implementation, which is never mentioned in the servlet.

Question: When I write a 2nd implementation class for the same interface, how do I control which implementation is injected into the servlet?

I'd like to control this via configuration so I can use a different impl class in different run modes. Secondarily, it might be useful to be able to select the impl dynamically at runtime, e.g if I need to select among implementations based on some request parameter.