In such a configuration, the MPP software uses the speech servers in a round-robin way. For example, if you configure a Nuance Vocalizer to use MRCPV1 and a Nuance Recognizer server to use MRCPv2 with TLS, the MPP software uses the Nuance Vocalizer MRCPv1 server for the first call, the Nuance Recognizer MRCPv2 with TLS for the second call, and so on.