In WebSphere Voice Server, a sequence of recognition commands that allocate a recognition engine and return a unique identifier to identify the engine.