Service Mesh Question

After watching your webinar on Service Mesh I have a question. I assume the primary reason to use the Service Mesh deployment pattern instead of just an internal API Gateway cluster pattern to accomplish the same task is latency. If this assumption is correct, then do you have any performance data on the difference between the two deployment patterns. I think a demo of a multiple hop calls in both deployment patterns with performance statistics would be very helpful to those of use evaluating these patterns.

Thank you