What are the pros/cons of sending context state in WCF Message Headers vs. using one of the WCF Context Bindings? Both seem to be able to solve my problem of out-of-band context state.
All of the context data that I have can be serialized as string key-value pairs.
I'm currently leaning toward Message Headers as it seems like it might be more consistent in the future should I choose a binding that doesn't have a context variant.
Has anyone run into a deal-breaker with either style?