Question

It might be a silly question, but say I have a hughe message that I want to process with Camel. How will the number of steps in my route affect the memory usage? Does camel deep copy my message payload for every step in the route, even if the DSL-step only reads from the message or does it do something smart here?

Is it better to keep the route down and do things in a “hughe” bean for large messages or not?

This is an example route that does various things, but not changing the payload.

from("foo:bar")
  .log(..)
  .setProperty(..)
  .setHeader(..)
  .log(..)
  .choice()
    .when(simple(... ) ) 
      .log(..)
      .to(..)
    .when(simple(..))
      .log(..)
      .to(..)
  .end()

Answer 1

from my understanding, for a simple pipelined route like this, an Exchange is created containing the body once and passed along each step in the route. Other EIPs do cause the Exchange to be copied though (like multicast, wiretap, etc)…

as well, if you have steps along the route which interface with external resources which could result in any type of copy/clone/conversion/serialization of the body unnecessarily, then you might use something like the claim check pattern to reduce this.