SpeechFrame - Architecture - Comsys.netArchitecture
The SpeechFrame® platform is based on Java technology, enabling the greatest flexibility to customers. The platform internally is designed in several layers. At the top the customer can determine the applications, far removed from the technical implementation. The applications are written in XML variants like VoiceXML and CCXML. The system is able to interact with any backoffice system and supports many different interfaces. It is also able to handle all sorts of telephony standards. A logical overview of the SpeechFrame® platform is shown here:
Physical Layer
The SpeechFrame® physical layer (SPL) consists of platform Ethernet connections and telephony board TDM and/or Ethernet connections. Network bearers – Speech and Video – can be transported over the TDM connections. The TDM connectors are located on the telephony board in the chassis. Most common bearers are the speech or video channels related to ISDN and ISUP calls. Next to TDM, the bearers can be transported over UDP (RTP). These Ethernet connectors are located on the telephony board in the chassis. Most common are RTP streams related to SIP calls.
Control Nodes & Signaling Nodes
The control nodes and signaling nodes use 3rd party API’s control the device drivers for the SpeechFrame® Physical Layer. The nodes are low-level processes, responsible for specific tasks.
Messaging Control Node
The Messaging Control Node (MCN) is responsible for handling non-call related messaging. Messages vary from SMS/MMS/USSD over TDM or SIGTRAN and SMS/MMS/USSD over SMPP/TCP network messaging to dialogue/service specific messaging. These types of messaging are connections to databases (i.e. JBDC), billing systems (i.e. TailorMade MD IVR Interface) and different kind of back-end interfaces based on SOAP/WSDL (i.e. Ericsson UCIP, SchlumbergerSema, COSI).
Service Signaling Node
The Service Signaling Node (SSN) is responsible for IN Signaling based on INAP (CS1/CH-IN2). While CS1+ and CS2 are not fully implemented, the SSN has several messages from CS1+ and CS2 implemented. The message set will be expanded on request.
Bearer Signaling Node
The Bearer Signaling Node (BSN) is responsible for all call-related signaling. It handles call control and call switching for the SpeechFrame platform. The BSN utilizes standard protocol signaling and is equipped to handle special signaling scenarios like, for example, ISUP call transfer based on special release cause and ISDN Explicit Call Transfer (ECT) route optimization & on-hold feature. SpeechFrame can be installed with Euro-ISDN (ETS300-102), ETSI-ISUP (ETS300-356) network protocols. For Voice over IP (VoIP), SIP is available and Media Resource Control Protocol (MRCP) is available for Speech Recognition and Text-To-Speech. Next to these protocols, numerous others can be delivered.
Bearer Control Node
The Bearer Control Node (BCN) is responsible for media-related control in the SpeechFrame platform. The node handles media resource allocation for playing, recording, conferencing, text-to-speech (TTS) and any other media demand. Besides media control, the BCN also controls bearer channel switching.
Signaling Logic & Control Logic
The SpeechFrame Signaling Logic (SSL) and SpeechFrame Control Logic (SCL) use Comsys proprietary TCP protocols to communicate with and control the MSN, SSN, BSN and BCN. The units exchange information on the same level based on incoming events and outgoing commands. Both the SSL and the SCL service the above building blocks, and are in fact the heart of the SpeechFrame® system, bridging the gap between low-level nodes and higher-level interpreters or java application.
|
