We propose a comprehensive model of an API-set that allows the developers to envision and realize a variety of distributed real-time multimedia applications as networks of easily understood and analyzable high-level real-time distributed computing objects, called the Time-triggered Message-triggered Objects (TMOs). This model reflects our view that any multimedia application can be represented by a set of TMOs networked properly. The software architecture and structuring model basically consists of four building-blocks which are the single-stream TMO, the compound stream TMO, the conductor TMO, and the stream connection. The definitions and roles of these building-blocks are discussed, and illustrated by applying them in structuring a multi-party audio-video teleconferencing service system.