I am examining your product for use in some high performance latency critical applications. I am aware of some small performance hit that occurs due to P/Invoke and managed calls via native code. Probably more concerning is the avoidance of the sudden high latency spikes that occur during garbage collection.
I'm assuming that given the nature of your product, you have seen this and have mechanism, strategies or ideas for how to accomplish latency critical serial communications from within managed code.
Can you please comment on, or offer any information as to how this is dealt with or if it is dealt with in your offering.
Essentially I'm looking for some indication that you have experience with this, and that your product will meet my needs [:)]