It isn’t all about software.
Video calling is notoriously tricky. Experience can be great when it works, but it can crumble really fast. Oftentimes, the reason for a bad experience is the network – be it ongoing congestion or packet loss. When you can’t get the media to the other end, media gets degraded.
This is the case for VoIP, video conferencing in the enterprise and telepresence. It is also the case for WebRTC. The following Avengers clip got me thinking exactly about this topic:
The takeaway?
- We are moving from managed networks towards the cloud
- It has been true forever to most services, but it is also becoming true for video services
- WebRTC accelerates this trend
- WebRTC services will win by designing and running world-class backend infrastructure that will reduce these problems, and it can be done
- The user experience needs to take into account such service glitches
Enjoy the clip.
I’m always surprised how badly these things work.
Even after so many years that someone hasn’t come up with a solution for this.
For example today I made a simple landline phonecall to someone else with a landline and still their were problems. It’s a pretty sad state when 2 phone companies can’t route some simple 64k or less voice call.
Anyway when making a video call, clearly there is enough bandwidth (temporarily). Can’t they just sent 2 audio streams, 1 with very low quality in the hopes at least 1 audio stream reaches the other end.
Lennie, while I do understand the frustration, I don’t think that sending two separate streams will work either – networks will typically kill and drop packets indiscriminately when they get congested.
There is no easy answer, and this is a tough problem to solve as there are many moving parts to deal with.