thanks for putting fuzzing into a wider context.

I still think it is important to fuzz for people who write their own RTP/RTCP implementations. It complements (hopefully) existing processes for code review (even though I can see how I would have approved the janus get_remb code) and unit testing. And I hope that i’ve shown it is not overly expensive or complicated so can be applied by all developers on your shortlist.
The business rationale for this is increased stability and uptime.

If we can make fuzzing common place in 2019 i’ll be happy. But I think we need to talk about it a bit more.