Can anybody tell me what the best practice is for creating and consuming multiple map layers in a dynamic service... my question centers around whether it is better to create one map service with dozens of different map layers contained within it, but then only make visible a few layers at a time, vs: creating multiple map services and then loading up the few that I need? I have multiple applications that will be consuming these services, but only a select few map services per application... What are the performance pros/cons to having all map layers in one service when only a few are being used in a given application? I'm guessing that results in a lighter load on the server than having multiple map services being called up from the server? The same question applies to geoprocessing services - is it better to have multiple geoprocessing tasks contained within one geoprocessing service, or would it be better to split out each unique task into its own geoprocessing service?
Thanks,
Ken