Robust Fault Handling
Lividi offers products and services for video conferencing and remote operations, where short latency is of paramount importance for interactivity and responsiveness. In order to provide robustness and resilience for theses challenging application requirements, a number of different techniques are utilized. These include:
The importance of different streams and parts of streams are adapted to application requirements. Stream priorities may also change dynamically. As an example, in case of video conferencing the audio stream from the current speaker might be most important. As another example, different packets for a single stream might also differ, as in case of scalable video (H.264/SVC) where the base layer is more important than enhancement layers. The differences in priorities allow adaptation by controlled dropping of the least important packets first, leading to graceful degradation to resource scarcity instead of failure.
The streamed media content, including video and audio, is protected by error correcting coding. Added redundancy is preferable over retransmissions in case of packet loss in order to provide the lowest possible end-to-end media delay.
A single network connection represents a single point of failure. This single point of failure renders the video conference session or remote operation session fragile. The Lividi multilink networking technology allows a number of networks to be used concurrently, for improved robustness and performance. As an example, a user may join a video conference session utilizing both wired (e.g., Ehternet) and wireless (e.g., WiFi, 3G) networks concurrently. When disconnecting the wired network, WiFi and 3G allow the session to continue, while multiple 3G connections from different network operators might be used when moving out of WiFi coverage. Session continuity is provided as long as at least one network connection is available at any time, irrespective of when that particular network connection became available.
Fail-over of media routing nodes
In any distributed system, the probability of a node failure increases as the number of nodes increases. The Lividi platform is designed to handle failures of media routing nodes, with only minimal impact on the application itself. The platform allows "hot standby" nodes, which in case of failure of one of the media routing nodes automatically replace the failed node.
In addition to fail-over, the Lividi platform allows re-routing to handle failure of nodes in a certain location. By re-routing, we refer to the ability of changing the path that media streams follow from media router to media router.