Some Context: We recently ran into an issue in our shop where we unexpectedly learned, at ArcGIS Enterprise 11.1, that shared services enable schema locking with no option to disable it. The result is that we had to change our service from shared to dedicated, fairly minor change that took us a bit of time to figure out since this is not readily apparent. Based on this community post (https://community.esri.com/t5/arcgis-enterprise-ideas/disable-schema-locking-on-shared-services/idc-... we were relieved to find out that shared services will have the option to disable schema locking at the next LTS ArcGIS Enterprise 11.3. However, that option is global; meaning all shared services must either be schema locking or not schema locking within your enterprise.
The Idea: Instead of requiring a deployment to put all their metaphorical eggs in one basket, and because this is a binary choice, this idea is to implement two pools of shared services. One that has schema locking enabled and the other pool with schema locking disabled. This will allow organizations the flexibility to put schema-sensitive applications (like field editing) into a shared pool and schema-insensitive applications (hot take: everything else) into a separate shared pool. The other benefit, is that it will be readily apparent which shared services are schema-locked and which ones are not. This way no one has to make the hard choice to implement a one-sized fits all approach when there are two legitimate, concurrent strategies that an enterprise should consider; shared services for all!
@TanuHoque