> 1. Does camel need ActiveMQ broker to run or it can work without ActiveMQ also ?
ActiveMQ is only one among many  components supported by Camel. So yes, you can run Camel without AMQ.
> 2. What are the possible use cases where I can use camel with ActiveMQ ?
Well, actually whenever you want to receive/send message to JMS you should consider using Camel :) . JMS usage usually involves using Enterprise Integration Patterns . And whenever you need to use EIP, you should consider using Camel for that purpose.
Component is a factory of the endpoints. And endpoints are used to send/receive messages using various protocols and data sources. In the JMS regards broker is something external for Camel. Camel is a kind of powerful JMS client. :)
Check out these  two books (written by our fellow committers) if you are looking for good introduction to Camel.
>Well, actually whenever you want to receive/send message to JMS you should consider using Camel :) . JMS >usage usually involves using Enterprise Integration Patterns . And whenever you need to use EIP, you >should consider using Camel for that purpose.
But ActiveMQ alone can send/receive JMS message. What specific camel will do in this case ?
> Do people normally run ActiveMQ within Karaf or outside of the container?
Usually you would like to put broker as a separated container for many
reasons - scalability, high availability and so forth. Having embedded
broker means also that the systems communicating via that broker are
tightly coupled - you won't be able to enqueue the messages while your
application is down (as the broker will be down at the same time).
My opinionated recommendation is to have a standalone broker by
default and stick to embedded one only if you really got a reason to