The setup: Two organizations want to edit replicas of the same address data through a #GeoData Service. Their edits occur locally throughout the day in versioned feature classes, they both run scripts to reconcile and post edits to their Default versions in the evening, and the deltas are synched through the replicas by the next morning. Both organizations use the #Attribute Assistant including the #GenerateID table.
Question: Does the GenerateID table have to be part of the replica? If so, how are duplicate IDs avoided since each set of editors is seeing the same numbers in a synched up replica? If it does not have to be part of the replica, should one of the organizations set their GenerateID table to have higher values to avoid future duplicates?