An application switch instantiates two application-side network service instances for the same application. Each network service instance is characterized by a common Internet Protocol (IP) address, a common Open Systems Interconnection (OSI) reference model layer 2 (L2) media access control (MAC) address, and a unique (for the application) supplemental L2 identifier. The application switch maintains a mapping between a {client IP address, client port} tuple and a particular instantiated network service instance based at least in part on the supplemental L2 identifier of a particular one of the instantiated first and second network service instances. When the application switch receives a client communication via an application switch client-side network, the application switch determines the particular instantiated network service instance corresponding to the { , } tuple based on the mapping, and switches the received client communication to the determined application-side network service instance.