SSP interfaces the world in 3 major directions - telephony, MRCP for ASR/TTS and management/monitoring interface.

​

SSP uses integrated SIP/RTP stacks, which are developed internally. Design of those stacks is streamlined, which improves performance. In testing 70 channels of SIP/VXML were tested on 1.8 GHz PC. RTP stack also provides a number of configuration variables for tuning for optimal performance and audio quality depending on the network parameters. Other SSP components also provide a large number of configuration parameters, which allows SSP to run on a wide spectrum of systems. Even older systems, like 300-500 MHz, will allow for use of 20-40 channels. Up to 500 channels can be run on newer hardware, depending on the applciation and features used.

​

SSP executes ASR and TTS operations using the MRCP protocol (V1 and V2). This W3C draft standard protocol is already supported by several vendors, and is expected to become the dominant access method in the future. In theory, any standard MRCP servers should work equally; in practice there appear to be minor implementation differences between vendors. For SSP, initial testing has been performed with Nuance, IBM WebSphere Voice Server, Lumenvox and Telisma ASR/TTS MRCP servers. We have also extensively tested SSP with our own MRCP server. In the future, other vendor’s servers will be tested as well. We are also working to provide alternative ASR/TTS integrations to provide a cost effective all-in-one solution.

​

SSP comes with a management/monitoring interface and a client program, which uses it to allow for run time access. This program is developed on MS Windows and is available in source form for OEM integrations.​