Since query layers bypass the geodatabase functionality, do they perform faster as compared to geodatabase feature classes when published as a mapservice?
I know this could actually be tricky because, it depends on complexity of the query used to build the query layer. So, assuming that only two to three tables are used to be build the query layer with a simple join based on a key, will it actually perform better than a geodatabase feature class (with the columns pre-added to it by processing) ?
