Yes. By assigning all servers in different racks to the same LAN identifier in SlapOS Master and by sharing the same switch in different racks, all servers in different racks are viewed as a single cluster resource which is managed in a uniform manner.
This approach is scalable through hypercube cabling. By connecting the second interface of each server to the switch of another rack, and by letting Rapid.Space's SDN (based on babel) discover shortest routes, the number of racks which are interconnected can be very high and reach thousands of servers, all part of the same "logical LAN".